[2025-10-21 03:06:23,605] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:687] baseline 0.000GB (+0.000GB allocated, +0.002GB reserved)
[2025-10-21 03:06:23,606] [INFO] [axolotl.cli.config.load_cfg:248] [PID:687] config:
{
  "activation_offloading": false,
  "adapter": "qlora",
  "axolotl_config_path": "config.yaml",
  "base_model": "Qwen/Qwen3-1.7B-Base",
  "base_model_config": "Qwen/Qwen3-1.7B-Base",
  "batch_size": 16,
  "bf16": false,
  "capabilities": {
    "bf16": true,
    "compute_capability": "sm_75",
    "fp8": false,
    "n_gpu": 2,
    "n_node": 1
  },
  "chat_template": "qwen3",
  "context_parallel_size": 1,
  "cut_cross_entropy": true,
  "dataloader_num_workers": 2,
  "dataloader_pin_memory": true,
  "dataloader_prefetch_factor": 8,
  "dataset_num_proc": 4,
  "dataset_prepared_path": "last_run_prepared",
  "datasets": [
    {
      "chat_template": "tokenizer_default",
      "message_property_mappings": {
        "content": "content",
        "role": "role"
      },
      "path": "Tung177/km-vi-translation",
      "roles_to_train": [
        "assistant"
      ],
      "split": "train",
      "trust_remote_code": false,
      "type": "chat_template"
    }
  ],
  "ddp": true,
  "device": "cuda:0",
  "device_map": {
    "": 0
  },
  "dion_rank_fraction": 1.0,
  "dion_rank_multiple_of": 1,
  "embeddings_skip_upcast": true,
  "env_capabilities": {
    "torch_version": "2.6.0"
  },
  "eval_batch_size": 8,
  "eval_causal_lm_metrics": [
    "sacrebleu",
    "comet",
    "ter",
    "chrf"
  ],
  "eval_max_new_tokens": 128,
  "eval_sample_packing": true,
  "eval_table_size": 0,
  "experimental_skip_move_to_device": true,
  "flash_attention": false,
  "fp16": true,
  "gradient_accumulation_steps": 1,
  "gradient_checkpointing": true,
  "gradient_checkpointing_kwargs": {
    "use_reentrant": false
  },
  "include_tkps": true,
  "learning_rate": 0.00018,
  "liger_fused_linear_cross_entropy": true,
  "liger_glu_activation": true,
  "liger_layer_norm": true,
  "liger_rms_norm": true,
  "liger_rope": true,
  "lisa_layers_attribute": "model.layers",
  "load_best_model_at_end": false,
  "load_in_4bit": true,
  "load_in_8bit": false,
  "local_rank": 0,
  "logging_steps": 1,
  "lora_alpha": 32,
  "lora_dropout": 0.0,
  "lora_mlp_kernel": true,
  "lora_o_kernel": true,
  "lora_qkv_kernel": true,
  "lora_r": 32,
  "lora_target_linear": true,
  "loraplus_lr_embedding": 1e-06,
  "lr_scheduler": "cosine_with_min_lr",
  "lr_scheduler_kwargs": {
    "min_lr_rate": 0.1
  },
  "max_grad_norm": 1.0,
  "mean_resizing_embeddings": false,
  "micro_batch_size": 8,
  "model_config_type": "qwen3",
  "num_epochs": 1.0,
  "optimizer": "adamw_8bit",
  "output_dir": "/kaggle/working/outputs/qwen-sft-mt-kmvi",
  "pad_to_sequence_len": true,
  "plugins": [
    "axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin",
    "axolotl.integrations.liger.LigerPlugin"
  ],
  "pretrain_multipack_attn": true,
  "profiler_steps_start": 0,
  "qlora_sharded_model_loading": false,
  "ray_num_workers": 1,
  "resources_per_worker": {
    "GPU": 1
  },
  "sample_packing": true,
  "sample_packing_bin_size": 200,
  "sample_packing_group_size": 100000,
  "save_only_model": false,
  "save_safetensors": true,
  "save_total_limit": 2,
  "saves_per_epoch": 1,
  "sequence_len": 4096,
  "shuffle_before_merging_datasets": false,
  "shuffle_merged_datasets": true,
  "skip_prepare_dataset": false,
  "streaming_multipack_buffer_size": 10000,
  "strict": false,
  "tensor_parallel_size": 1,
  "tiled_mlp_use_original_mlp": true,
  "tokenizer_config": "Qwen/Qwen3-1.7B-Base",
  "tokenizer_save_jinja_files": true,
  "torch_dtype": "torch.float16",
  "train_on_inputs": false,
  "trl": {
    "log_completions": false,
    "mask_truncated_completions": false,
    "ref_model_mixup_alpha": 0.9,
    "ref_model_sync_steps": 64,
    "scale_rewards": true,
    "sync_ref_model": false,
    "use_vllm": false,
    "vllm_server_host": "0.0.0.0",
    "vllm_server_port": 8000
  },
  "use_ray": false,
  "use_wandb": true,
  "val_set_size": 0.0,
  "vllm": {
    "device": "auto",
    "dtype": "auto",
    "gpu_memory_utilization": 0.9,
    "host": "0.0.0.0",
    "port": 8000
  },
  "wandb_log_model": "false",
  "wandb_name": "qwen3-1.7B-mt-kmvi-v0.2",
  "wandb_project": "MT-V1",
  "warmup_ratio": 0.03,
  "weight_decay": 0.01,
  "world_size": 2,
  "xformers_attention": true
}
[2025-10-21 03:06:24,328] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:687] EOS: 151643 / <|endoftext|>
[2025-10-21 03:06:24,328] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:687] BOS: None / None
[2025-10-21 03:06:24,328] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:687] PAD: 151643 / <|endoftext|>
[2025-10-21 03:06:24,328] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:687] UNK: None / None
[2025-10-21 03:06:24,329] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:470] [PID:687] Loading prepared dataset from disk at last_run_prepared/85b14c2eb729e995064f439e6a0d9ea3...
[2025-10-21 03:06:25,136] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:406] [PID:687] total_num_tokens: 80_388_735
[2025-10-21 03:06:30,179] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:424] [PID:687] `total_supervised_tokens: 13_464_940`
[2025-10-21 03:06:37,446] [DEBUG] [axolotl.utils.samplers.multipack.pack_parallel:177] [PID:687] Using single process for pack_parallel, running sequentially.
[2025-10-21 03:06:39,947] [DEBUG] [axolotl.utils.samplers.multipack.pack_parallel:177] [PID:687] Using single process for pack_parallel, running sequentially.
[2025-10-21 03:06:41,103] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.2863664627075195
[2025-10-21 03:06:41,232] [DEBUG] [axolotl.utils.samplers.multipack.pack_parallel:177] [PID:687] Using single process for pack_parallel, running sequentially.
[2025-10-21 03:06:42,396] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.285015344619751
[2025-10-21 03:06:42,529] [DEBUG] [axolotl.utils.samplers.multipack.pack_parallel:177] [PID:687] Using single process for pack_parallel, running sequentially.
[2025-10-21 03:06:43,686] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.2829370498657227
[2025-10-21 03:06:43,815] [DEBUG] [axolotl.utils.samplers.multipack.pack_parallel:177] [PID:687] Using single process for pack_parallel, running sequentially.
[2025-10-21 03:06:44,982] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.288405179977417
[2025-10-21 03:06:45,908] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:438] [PID:687] gather_len_batches: [2460, 2460]
[2025-10-21 03:06:45,931] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:483] [PID:687] data_loader_len: 1230
[2025-10-21 03:06:45,954] [INFO] [axolotl.utils.trainer.calc_sample_packing_eff_est:499] [PID:687] sample_packing_eff_est across ranks: [0.9968081712722778, 0.9968081712722778]
[2025-10-21 03:06:45,955] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:511] [PID:687] sample_packing_eff_est: 1.0
[2025-10-21 03:06:45,955] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:522] [PID:687] total_num_steps: 1230
[2025-10-21 03:06:45,963] [INFO] [axolotl.utils.data.sft._prepare_standard_dataset:121] [PID:687] Maximum number of steps set at 1230
[2025-10-21 03:06:46,005] [DEBUG] [axolotl.train.setup_model_and_tokenizer:65] [PID:687] Loading tokenizer... Qwen/Qwen3-1.7B-Base
[2025-10-21 03:06:46,620] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:687] EOS: 151643 / <|endoftext|>
[2025-10-21 03:06:46,620] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:687] BOS: None / None
[2025-10-21 03:06:46,620] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:687] PAD: 151643 / <|endoftext|>
[2025-10-21 03:06:46,620] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:687] UNK: None / None
[2025-10-21 03:06:46,620] [DEBUG] [axolotl.train.setup_model_and_tokenizer:74] [PID:687] Loading model
[2025-10-21 03:06:46,948] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:687] Patched Trainer.evaluation_loop with nanmean loss calculation
[2025-10-21 03:06:46,949] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:687] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation
[2025-10-21 03:06:47,057] [INFO] [axolotl.monkeypatch.peft.utils.patch_peft_prep_code:74] [PID:687] patching prepare_model_for_kbit_training to allow for overrides
[2025-10-21 03:06:47,058] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:301] [PID:687] Applying multipack dataloader patch for sample packing...
[2025-10-21 03:06:47,161] [INFO] [axolotl.monkeypatch.lora_kernels.patch_self_attn_lora:237] [PID:687] Patched attention class with LoRA optims: Qwen3Attention
[2025-10-21 03:06:48,540] [INFO] [axolotl.integrations.cut_cross_entropy.pre_model_load:94] [PID:687] Applying Cut Cross Entropy to model type: qwen3
[2025-10-21 03:06:48,563] [INFO] [axolotl.integrations.liger.plugin.pre_model_load:71] [PID:687] Applying LIGER to qwen3 with kwargs: {'rope': True, 'cross_entropy': None, 'fused_linear_cross_entropy': True, 'rms_norm': True, 'swiglu': True}
[2025-10-21 03:07:15,601] [INFO] [axolotl.loaders.model._prepare_model_for_quantization:851] [PID:687] converting PEFT model w/ prepare_model_for_kbit_training
[2025-10-21 03:07:15,606] [INFO] [axolotl.loaders.model._configure_embedding_dtypes:345] [PID:687] Converting modules to torch.float16
[2025-10-21 03:07:15,610] [DEBUG] [axolotl.loaders.model.log_gpu_memory_usage:127] [PID:687] Memory usage after model load 1.314GB (+1.314GB allocated, +1.328GB reserved)
[2025-10-21 03:07:15,610] [INFO] [axolotl.loaders.adapter.load_lora:80] [PID:687] found linear modules: ['down_proj', 'gate_proj', 'k_proj', 'o_proj', 'q_proj', 'up_proj', 'v_proj']
trainable params: 34,865,152 || all params: 1,755,440,128 || trainable%: 1.9861
[2025-10-21 03:07:16,347] [DEBUG] [axolotl.loaders.model.log_gpu_memory_usage:127] [PID:687] after adapters 1.402GB (+1.402GB allocated, +1.457GB reserved)
[2025-10-21 03:07:33,541] [INFO] [axolotl.train.save_initial_configs:398] [PID:687] Pre-saving adapter config to /kaggle/working/outputs/qwen-sft-mt-kmvi...
[2025-10-21 03:07:33,542] [INFO] [axolotl.train.save_initial_configs:402] [PID:687] Pre-saving tokenizer to /kaggle/working/outputs/qwen-sft-mt-kmvi...
[2025-10-21 03:07:33,741] [INFO] [axolotl.train.save_initial_configs:407] [PID:687] Pre-saving model config to /kaggle/working/outputs/qwen-sft-mt-kmvi...
[2025-10-21 03:07:33,744] [INFO] [axolotl.train.execute_training:196] [PID:687] Starting trainer...
[2025-10-21 03:08:24,057] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.52178955078125
[2025-10-21 03:08:25,663] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.5997445583343506
[2025-10-21 03:08:27,235] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.5644276142120361
[2025-10-21 03:08:28,612] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:687] generate_batches time: 1.3716719150543213
[2025-10-21 03:08:28,614] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:438] [PID:687] gather_len_batches: [2827, 2827]
wandb: Currently logged in as: hasontung to https://api.wandb.ai. Use `wandb login --relogin` to force relogin
wandb: Tracking run with wandb version 0.21.0
wandb: Run data is saved locally in /kaggle/working/axolotl/wandb/run-20251021_030829-zoptfzaf
wandb: Run `wandb offline` to turn off syncing.
wandb: Syncing run qwen3-1.7B-mt-kmvi-v0.2
wandb: ⭐️ View project at https://wandb.ai/hasontung/MT-V1
wandb: 🚀 View run at https://wandb.ai/hasontung/MT-V1/runs/zoptfzaf
wandb: WARNING Saving files without folders. If you want to preserve subdirectories pass base_path to wandb.save, i.e. wandb.save("/mnt/folder/file.h5", base_path="/mnt")
[2025-10-21 03:08:31,397] [INFO] [axolotl.utils.callbacks.on_train_begin:757] [PID:687] The Axolotl config has been saved to the WandB run under files.

  0%|                                                  | 0/1230 [00:00<?, ?it/s]
  0%|                                       | 1/1230 [00:31<10:38:17, 31.16s/it]
                                                                                
{'loss': 2.1835, 'grad_norm': 1.4715336561203003, 'learning_rate': 0.0, 'memory/max_active (GiB)': 10.83, 'memory/max_allocated (GiB)': 10.83, 'memory/device_reserved (GiB)': 11.19, 'tokens_per_second_per_gpu': 169.41, 'epoch': 0.0}

  0%|                                       | 1/1230 [00:31<10:38:17, 31.16s/it]
  0%|                                        | 2/1230 [00:57<9:37:31, 28.22s/it]
                                                                                
{'loss': 2.2983, 'grad_norm': 1.5340197086334229, 'learning_rate': 5e-06, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 11.42, 'tokens_per_second_per_gpu': 193.94, 'epoch': 0.0}

  0%|                                        | 2/1230 [00:57<9:37:31, 28.22s/it]
  0%|                                        | 3/1230 [01:23<9:21:56, 27.48s/it]
                                                                                
{'loss': 2.2312, 'grad_norm': 1.4970414638519287, 'learning_rate': 1e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 11.42, 'tokens_per_second_per_gpu': 170.87, 'epoch': 0.0}

  0%|                                        | 3/1230 [01:23<9:21:56, 27.48s/it]
  0%|▏                                       | 4/1230 [01:49<9:09:15, 26.88s/it]
                                                                                
{'loss': 2.203, 'grad_norm': 1.5545872449874878, 'learning_rate': 1.5e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 187.19, 'epoch': 0.0}

  0%|▏                                       | 4/1230 [01:49<9:09:15, 26.88s/it]
  0%|▏                                       | 5/1230 [02:12<8:39:55, 25.47s/it]
                                                                                
{'loss': 2.2044, 'grad_norm': 1.6360639333724976, 'learning_rate': 2e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 198.2, 'epoch': 0.0}

  0%|▏                                       | 5/1230 [02:12<8:39:55, 25.47s/it]
  0%|▏                                       | 6/1230 [02:36<8:25:19, 24.77s/it]
                                                                                
{'loss': 2.2369, 'grad_norm': 1.5867512226104736, 'learning_rate': 2.5e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 200.07, 'epoch': 0.0}

  0%|▏                                       | 6/1230 [02:36<8:25:19, 24.77s/it]
  1%|▏                                       | 7/1230 [03:03<8:40:04, 25.51s/it]
                                                                                
{'loss': 2.0881, 'grad_norm': 1.5065139532089233, 'learning_rate': 3e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 180.53, 'epoch': 0.0}

  1%|▏                                       | 7/1230 [03:03<8:40:04, 25.51s/it]
  1%|▎                                       | 8/1230 [03:27<8:28:17, 24.96s/it]
                                                                                
{'loss': 1.9851, 'grad_norm': 1.4327116012573242, 'learning_rate': 3.5000000000000004e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 200.48, 'epoch': 0.01}

  1%|▎                                       | 8/1230 [03:27<8:28:17, 24.96s/it]
  1%|▎                                       | 9/1230 [03:55<8:47:45, 25.93s/it]
                                                                                
{'loss': 1.9712, 'grad_norm': 1.2687053680419922, 'learning_rate': 4e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 162.53, 'epoch': 0.01}

  1%|▎                                       | 9/1230 [03:55<8:47:45, 25.93s/it]
  1%|▎                                      | 10/1230 [04:23<9:03:26, 26.73s/it]
                                                                                
{'loss': 1.8343, 'grad_norm': 1.0051155090332031, 'learning_rate': 4.5e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 172.25, 'epoch': 0.01}

  1%|▎                                      | 10/1230 [04:23<9:03:26, 26.73s/it]
  1%|▎                                      | 11/1230 [04:52<9:14:05, 27.27s/it]
                                                                                
{'loss': 1.8082, 'grad_norm': 0.9459324479103088, 'learning_rate': 5e-05, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 181.91, 'epoch': 0.01}

  1%|▎                                      | 11/1230 [04:52<9:14:05, 27.27s/it]
  1%|▍                                      | 12/1230 [05:20<9:22:56, 27.73s/it]
                                                                                
{'loss': 1.753, 'grad_norm': 0.9601195454597473, 'learning_rate': 5.500000000000001e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 153.89, 'epoch': 0.01}

  1%|▍                                      | 12/1230 [05:20<9:22:56, 27.73s/it]
  1%|▍                                      | 13/1230 [05:47<9:15:12, 27.37s/it]
                                                                                
{'loss': 1.7189, 'grad_norm': 0.8818406462669373, 'learning_rate': 6e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 181.22, 'epoch': 0.01}

  1%|▍                                      | 13/1230 [05:47<9:15:12, 27.37s/it]
  1%|▍                                      | 14/1230 [06:16<9:25:59, 27.93s/it]
                                                                                
{'loss': 1.6157, 'grad_norm': 0.6909472942352295, 'learning_rate': 6.500000000000001e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 160.36, 'epoch': 0.01}

  1%|▍                                      | 14/1230 [06:16<9:25:59, 27.93s/it]
  1%|▍                                      | 15/1230 [06:41<9:06:44, 27.00s/it]
                                                                                
{'loss': 1.5687, 'grad_norm': 0.566254734992981, 'learning_rate': 7.000000000000001e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 188.93, 'epoch': 0.01}

  1%|▍                                      | 15/1230 [06:41<9:06:44, 27.00s/it]
  1%|▌                                      | 16/1230 [07:06<8:54:30, 26.42s/it]
                                                                                
{'loss': 1.5225, 'grad_norm': 0.4967886507511139, 'learning_rate': 7.500000000000001e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 189.91, 'epoch': 0.01}

  1%|▌                                      | 16/1230 [07:06<8:54:30, 26.42s/it]
  1%|▌                                      | 17/1230 [07:36<9:13:44, 27.39s/it]
                                                                                
{'loss': 1.5403, 'grad_norm': 0.4461193084716797, 'learning_rate': 8e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 158.24, 'epoch': 0.01}

  1%|▌                                      | 17/1230 [07:36<9:13:44, 27.39s/it]
  1%|▌                                      | 18/1230 [08:01<9:01:33, 26.81s/it]
                                                                                
{'loss': 1.435, 'grad_norm': 0.4377662241458893, 'learning_rate': 8.5e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 177.89, 'epoch': 0.01}

  1%|▌                                      | 18/1230 [08:01<9:01:33, 26.81s/it]
  2%|▌                                      | 19/1230 [08:31<9:19:55, 27.74s/it]
                                                                                
{'loss': 1.4565, 'grad_norm': 0.43463006615638733, 'learning_rate': 9e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 161.54, 'epoch': 0.01}

  2%|▌                                      | 19/1230 [08:31<9:19:55, 27.74s/it]
  2%|▋                                      | 20/1230 [09:01<9:31:41, 28.35s/it]
                                                                                
{'loss': 1.4844, 'grad_norm': 0.435220867395401, 'learning_rate': 9.5e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 147.09, 'epoch': 0.01}

  2%|▋                                      | 20/1230 [09:01<9:31:41, 28.35s/it]
  2%|▋                                      | 21/1230 [09:26<9:09:45, 27.28s/it]
                                                                                
{'loss': 1.4109, 'grad_norm': 0.4355340003967285, 'learning_rate': 0.0001, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 182.6, 'epoch': 0.01}

  2%|▋                                      | 21/1230 [09:26<9:09:45, 27.28s/it]
  2%|▋                                      | 22/1230 [09:55<9:22:04, 27.92s/it]
                                                                                
{'loss': 1.4855, 'grad_norm': 0.42590785026550293, 'learning_rate': 0.00010500000000000002, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 165.52, 'epoch': 0.02}

  2%|▋                                      | 22/1230 [09:55<9:22:04, 27.92s/it]
  2%|▋                                      | 23/1230 [10:25<9:30:57, 28.38s/it]
                                                                                
{'loss': 1.3763, 'grad_norm': 0.43551531434059143, 'learning_rate': 0.00011000000000000002, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 150.12, 'epoch': 0.02}

  2%|▋                                      | 23/1230 [10:25<9:30:57, 28.38s/it]
  2%|▊                                      | 24/1230 [10:50<9:13:15, 27.52s/it]
                                                                                
{'loss': 1.4209, 'grad_norm': 0.4493299424648285, 'learning_rate': 0.000115, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 176.47, 'epoch': 0.02}

  2%|▊                                      | 24/1230 [10:50<9:13:15, 27.52s/it]
  2%|▊                                      | 25/1230 [11:20<9:26:50, 28.22s/it]
                                                                                
{'loss': 1.4178, 'grad_norm': 0.420244425535202, 'learning_rate': 0.00012, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 154.78, 'epoch': 0.02}

  2%|▊                                      | 25/1230 [11:20<9:26:50, 28.22s/it]
  2%|▊                                      | 26/1230 [11:46<9:11:21, 27.48s/it]
                                                                                
{'loss': 1.3736, 'grad_norm': 0.43762001395225525, 'learning_rate': 0.000125, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 180.17, 'epoch': 0.02}

  2%|▊                                      | 26/1230 [11:46<9:11:21, 27.48s/it]
  2%|▊                                      | 27/1230 [12:16<9:25:57, 28.23s/it]
                                                                                
{'loss': 1.3282, 'grad_norm': 0.39991483092308044, 'learning_rate': 0.00013000000000000002, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 157.0, 'epoch': 0.02}

  2%|▊                                      | 27/1230 [12:16<9:25:57, 28.23s/it]
  2%|▉                                      | 28/1230 [12:45<9:34:27, 28.68s/it]
                                                                                
{'loss': 1.3026, 'grad_norm': 0.4060404300689697, 'learning_rate': 0.000135, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 155.09, 'epoch': 0.02}

  2%|▉                                      | 28/1230 [12:45<9:34:27, 28.68s/it]
  2%|▉                                      | 29/1230 [13:11<9:14:05, 27.68s/it]
                                                                                
{'loss': 1.3046, 'grad_norm': 0.3773941397666931, 'learning_rate': 0.00014000000000000001, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 189.76, 'epoch': 0.02}

  2%|▉                                      | 29/1230 [13:11<9:14:05, 27.68s/it]
  2%|▉                                      | 30/1230 [13:36<8:59:59, 27.00s/it]
                                                                                
{'loss': 1.3569, 'grad_norm': 0.3649119436740875, 'learning_rate': 0.000145, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.46, 'epoch': 0.02}

  2%|▉                                      | 30/1230 [13:36<8:59:59, 27.00s/it]
  3%|▉                                      | 31/1230 [14:01<8:49:11, 26.48s/it]
                                                                                
{'loss': 1.3145, 'grad_norm': 0.38387608528137207, 'learning_rate': 0.00015000000000000001, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 192.29, 'epoch': 0.02}

  3%|▉                                      | 31/1230 [14:01<8:49:11, 26.48s/it]
  3%|█                                      | 32/1230 [14:31<9:06:54, 27.39s/it]
                                                                                
{'loss': 1.302, 'grad_norm': 0.3419938087463379, 'learning_rate': 0.00015500000000000003, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 163.99, 'epoch': 0.02}

  3%|█                                      | 32/1230 [14:31<9:06:54, 27.39s/it]
  3%|█                                      | 33/1230 [15:01<9:20:23, 28.09s/it]
                                                                                
{'loss': 1.3192, 'grad_norm': 0.37636736035346985, 'learning_rate': 0.00016, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 178.24, 'epoch': 0.02}

  3%|█                                      | 33/1230 [15:01<9:20:23, 28.09s/it]
  3%|█                                      | 34/1230 [15:26<9:02:16, 27.20s/it]
                                                                                
{'loss': 1.2029, 'grad_norm': 0.3419751524925232, 'learning_rate': 0.000165, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.87, 'epoch': 0.02}

  3%|█                                      | 34/1230 [15:26<9:02:16, 27.20s/it]
  3%|█                                      | 35/1230 [15:55<9:15:18, 27.88s/it]
                                                                                
{'loss': 1.2349, 'grad_norm': 0.3449782729148865, 'learning_rate': 0.00017, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 157.71, 'epoch': 0.02}

  3%|█                                      | 35/1230 [15:55<9:15:18, 27.88s/it]
  3%|█▏                                     | 36/1230 [16:21<9:04:02, 27.34s/it]
                                                                                
{'loss': 1.2412, 'grad_norm': 0.36126509308815, 'learning_rate': 0.000175, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 187.27, 'epoch': 0.03}

  3%|█▏                                     | 36/1230 [16:21<9:04:02, 27.34s/it]
  3%|█▏                                     | 37/1230 [16:47<8:52:13, 26.77s/it]
                                                                                
{'loss': 1.2391, 'grad_norm': 0.34673699736595154, 'learning_rate': 0.00018, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 181.45, 'epoch': 0.03}

  3%|█▏                                     | 37/1230 [16:47<8:52:13, 26.77s/it]
  3%|█▏                                     | 38/1230 [17:12<8:45:32, 26.45s/it]
                                                                                
{'loss': 1.2803, 'grad_norm': 0.351216584444046, 'learning_rate': 0.0001799997196207534, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 183.5, 'epoch': 0.03}

  3%|█▏                                     | 38/1230 [17:12<8:45:32, 26.45s/it]
  3%|█▏                                     | 39/1230 [17:39<8:43:51, 26.39s/it]
                                                                                
{'loss': 1.2081, 'grad_norm': 0.35957658290863037, 'learning_rate': 0.00017999887848495455, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 186.24, 'epoch': 0.03}

  3%|█▏                                     | 39/1230 [17:39<8:43:51, 26.39s/it]
  3%|█▎                                     | 40/1230 [18:04<8:37:41, 26.10s/it]
                                                                                
{'loss': 1.2231, 'grad_norm': 0.3385028839111328, 'learning_rate': 0.00017999747659842668, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 178.7, 'epoch': 0.03}

  3%|█▎                                     | 40/1230 [18:04<8:37:41, 26.10s/it]
  3%|█▎                                     | 41/1230 [18:30<8:33:04, 25.89s/it]
                                                                                
{'loss': 1.2341, 'grad_norm': 0.3493475914001465, 'learning_rate': 0.0001799955139708749, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 183.79, 'epoch': 0.03}

  3%|█▎                                     | 41/1230 [18:30<8:33:04, 25.89s/it]
  3%|█▎                                     | 42/1230 [18:55<8:31:46, 25.85s/it]
                                                                                
{'loss': 1.1889, 'grad_norm': 0.35491707921028137, 'learning_rate': 0.0001799929906158864, 'memory/max_active (GiB)': 7.97, 'memory/max_allocated (GiB)': 7.97, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 212.73, 'epoch': 0.03}

  3%|█▎                                     | 42/1230 [18:55<8:31:46, 25.85s/it]
  3%|█▎                                     | 43/1230 [19:25<8:54:50, 27.03s/it]
                                                                                
{'loss': 1.1984, 'grad_norm': 0.3413371443748474, 'learning_rate': 0.00017998990655093025, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 163.62, 'epoch': 0.03}

  3%|█▎                                     | 43/1230 [19:25<8:54:50, 27.03s/it]
  4%|█▍                                     | 44/1230 [19:55<9:09:36, 27.80s/it]
                                                                                
{'loss': 1.1954, 'grad_norm': 0.35759031772613525, 'learning_rate': 0.00017998626179735724, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 154.67, 'epoch': 0.03}

  4%|█▍                                     | 44/1230 [19:55<9:09:36, 27.80s/it]
  4%|█▍                                     | 45/1230 [20:21<8:57:29, 27.22s/it]
                                                                                
{'loss': 1.17, 'grad_norm': 0.32557523250579834, 'learning_rate': 0.00017998205638039978, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 190.41, 'epoch': 0.03}

  4%|█▍                                     | 45/1230 [20:21<8:57:29, 27.22s/it]
  4%|█▍                                     | 46/1230 [20:50<9:08:13, 27.78s/it]
                                                                                
{'loss': 1.1404, 'grad_norm': 0.3352647125720978, 'learning_rate': 0.00017997729032917173, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 164.06, 'epoch': 0.03}

  4%|█▍                                     | 46/1230 [20:50<9:08:13, 27.78s/it]
  4%|█▍                                     | 47/1230 [21:15<8:52:51, 27.03s/it]
                                                                                
{'loss': 1.182, 'grad_norm': 0.36593785881996155, 'learning_rate': 0.00017997196367666824, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 182.99, 'epoch': 0.03}

  4%|█▍                                     | 47/1230 [21:15<8:52:51, 27.03s/it]
  4%|█▌                                     | 48/1230 [21:44<9:06:25, 27.74s/it]
                                                                                
{'loss': 1.1677, 'grad_norm': 0.3575313687324524, 'learning_rate': 0.00017996607645976538, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 158.24, 'epoch': 0.03}

  4%|█▌                                     | 48/1230 [21:44<9:06:25, 27.74s/it]
  4%|█▌                                     | 49/1230 [22:10<8:52:12, 27.04s/it]
                                                                                
{'loss': 1.1342, 'grad_norm': 0.35924556851387024, 'learning_rate': 0.0001799596287192201, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 178.83, 'epoch': 0.03}

  4%|█▌                                     | 49/1230 [22:10<8:52:12, 27.04s/it]
  4%|█▌                                     | 50/1230 [22:35<8:43:51, 26.64s/it]
                                                                                
{'loss': 1.2219, 'grad_norm': 0.3943169116973877, 'learning_rate': 0.00017995262049966967, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 200.65, 'epoch': 0.04}

  4%|█▌                                     | 50/1230 [22:35<8:43:51, 26.64s/it]
  4%|█▌                                     | 51/1230 [23:05<9:00:44, 27.52s/it]
                                                                                
{'loss': 1.1544, 'grad_norm': 0.36784133315086365, 'learning_rate': 0.00017994505184963164, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 160.16, 'epoch': 0.04}

  4%|█▌                                     | 51/1230 [23:05<9:00:44, 27.52s/it]
  4%|█▋                                     | 52/1230 [23:31<8:51:54, 27.09s/it]
                                                                                
{'loss': 1.1907, 'grad_norm': 0.38494810461997986, 'learning_rate': 0.00017993692282150336, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.68, 'epoch': 0.04}

  4%|█▋                                     | 52/1230 [23:31<8:51:54, 27.09s/it]
  4%|█▋                                     | 53/1230 [24:00<9:04:41, 27.77s/it]
                                                                                
{'loss': 1.0965, 'grad_norm': 0.3574252724647522, 'learning_rate': 0.00017992823347156157, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 156.4, 'epoch': 0.04}

  4%|█▋                                     | 53/1230 [24:00<9:04:41, 27.77s/it]
  4%|█▋                                     | 54/1230 [24:26<8:50:53, 27.09s/it]
                                                                                
{'loss': 1.1629, 'grad_norm': 0.3886204659938812, 'learning_rate': 0.00017991898385996227, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 179.68, 'epoch': 0.04}

  4%|█▋                                     | 54/1230 [24:26<8:50:53, 27.09s/it]
  4%|█▋                                     | 55/1230 [24:56<9:05:12, 27.84s/it]
                                                                                
{'loss': 1.1387, 'grad_norm': 0.3612476885318756, 'learning_rate': 0.00017990917405073993, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 163.27, 'epoch': 0.04}

  4%|█▋                                     | 55/1230 [24:56<9:05:12, 27.84s/it]
  5%|█▊                                     | 56/1230 [25:25<9:12:49, 28.25s/it]
                                                                                
{'loss': 1.1751, 'grad_norm': 0.36159127950668335, 'learning_rate': 0.00017989880411180734, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 156.38, 'epoch': 0.04}

  5%|█▊                                     | 56/1230 [25:25<9:12:49, 28.25s/it]
  5%|█▊                                     | 57/1230 [25:54<9:17:15, 28.50s/it]
                                                                                
{'loss': 1.1641, 'grad_norm': 0.3770463168621063, 'learning_rate': 0.000179887874114955, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 149.37, 'epoch': 0.04}

  5%|█▊                                     | 57/1230 [25:54<9:17:15, 28.50s/it]
  5%|█▊                                     | 58/1230 [26:19<8:58:31, 27.57s/it]
                                                                                
{'loss': 1.1257, 'grad_norm': 0.38490909337997437, 'learning_rate': 0.0001798763841358507, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 187.07, 'epoch': 0.04}

  5%|█▊                                     | 58/1230 [26:19<8:58:31, 27.57s/it]
  5%|█▊                                     | 59/1230 [26:45<8:45:43, 26.94s/it]
                                                                                
{'loss': 1.1538, 'grad_norm': 0.3650124967098236, 'learning_rate': 0.00017986433425403887, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 204.38, 'epoch': 0.04}

  5%|█▊                                     | 59/1230 [26:45<8:45:43, 26.94s/it]
  5%|█▉                                     | 60/1230 [27:10<8:35:05, 26.42s/it]
                                                                                
{'loss': 1.1236, 'grad_norm': 0.3694992661476135, 'learning_rate': 0.0001798517245529402, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 185.1, 'epoch': 0.04}

  5%|█▉                                     | 60/1230 [27:10<8:35:05, 26.42s/it]
  5%|█▉                                     | 61/1230 [27:39<8:52:06, 27.31s/it]
                                                                                
{'loss': 1.147, 'grad_norm': 0.36638060212135315, 'learning_rate': 0.00017983855511985097, 'memory/max_active (GiB)': 7.96, 'memory/max_allocated (GiB)': 7.96, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 183.08, 'epoch': 0.04}

  5%|█▉                                     | 61/1230 [27:39<8:52:06, 27.31s/it]
  5%|█▉                                     | 62/1230 [28:09<9:05:17, 28.01s/it]
                                                                                
{'loss': 1.1323, 'grad_norm': 0.38691216707229614, 'learning_rate': 0.0001798248260459424, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 155.5, 'epoch': 0.04}

  5%|█▉                                     | 62/1230 [28:09<9:05:17, 28.01s/it]
  5%|█▉                                     | 63/1230 [28:35<8:51:55, 27.35s/it]
                                                                                
{'loss': 1.1359, 'grad_norm': 0.3812416195869446, 'learning_rate': 0.00017981053742626013, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 197.78, 'epoch': 0.04}

  5%|█▉                                     | 63/1230 [28:35<8:51:55, 27.35s/it]
  5%|██                                     | 64/1230 [29:00<8:41:50, 26.85s/it]
                                                                                
{'loss': 1.0725, 'grad_norm': 0.3674742877483368, 'learning_rate': 0.00017979568935972342, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 204.46, 'epoch': 0.05}

  5%|██                                     | 64/1230 [29:00<8:41:50, 26.85s/it]
  5%|██                                     | 65/1230 [29:27<8:37:24, 26.65s/it]
                                                                                
{'loss': 1.1056, 'grad_norm': 0.382534384727478, 'learning_rate': 0.0001797802819491247, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 191.42, 'epoch': 0.05}

  5%|██                                     | 65/1230 [29:27<8:37:24, 26.65s/it]
  5%|██                                     | 66/1230 [29:52<8:27:19, 26.15s/it]
                                                                                
{'loss': 1.0782, 'grad_norm': 0.3882080316543579, 'learning_rate': 0.00017976431530112859, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 193.92, 'epoch': 0.05}

  5%|██                                     | 66/1230 [29:52<8:27:19, 26.15s/it]
  5%|██                                     | 67/1230 [30:17<8:22:12, 25.91s/it]
                                                                                
{'loss': 1.0919, 'grad_norm': 0.4063616394996643, 'learning_rate': 0.00017974778952627128, 'memory/max_active (GiB)': 9.76, 'memory/max_allocated (GiB)': 9.76, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 168.53, 'epoch': 0.05}

  5%|██                                     | 67/1230 [30:17<8:22:12, 25.91s/it]
  6%|██▏                                    | 68/1230 [30:42<8:17:30, 25.69s/it]
                                                                                
{'loss': 1.0636, 'grad_norm': 0.3898955285549164, 'learning_rate': 0.00017973070473895975, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.3, 'epoch': 0.05}

  6%|██▏                                    | 68/1230 [30:42<8:17:30, 25.69s/it]
  6%|██▏                                    | 69/1230 [31:08<8:18:58, 25.79s/it]
                                                                                
{'loss': 1.1015, 'grad_norm': 0.3781646192073822, 'learning_rate': 0.0001797130610574711, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 192.63, 'epoch': 0.05}

  6%|██▏                                    | 69/1230 [31:08<8:18:58, 25.79s/it]
  6%|██▏                                    | 70/1230 [31:38<8:41:18, 26.96s/it]
                                                                                
{'loss': 1.0753, 'grad_norm': 0.4182055592536926, 'learning_rate': 0.00017969485860395157, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 151.52, 'epoch': 0.05}

  6%|██▏                                    | 70/1230 [31:38<8:41:18, 26.96s/it]
  6%|██▎                                    | 71/1230 [32:07<8:55:28, 27.72s/it]
                                                                                
{'loss': 1.1067, 'grad_norm': 0.4031315743923187, 'learning_rate': 0.00017967609750441565, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 148.88, 'epoch': 0.05}

  6%|██▎                                    | 71/1230 [32:07<8:55:28, 27.72s/it]
  6%|██▎                                    | 72/1230 [32:33<8:43:37, 27.13s/it]
                                                                                
{'loss': 1.0948, 'grad_norm': 0.40818241238594055, 'learning_rate': 0.00017965677788874549, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 195.26, 'epoch': 0.05}

  6%|██▎                                    | 72/1230 [32:33<8:43:37, 27.13s/it]
  6%|██▎                                    | 73/1230 [32:59<8:38:01, 26.86s/it]
                                                                                
{'loss': 1.0895, 'grad_norm': 0.39298272132873535, 'learning_rate': 0.0001796368998906896, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 182.24, 'epoch': 0.05}

  6%|██▎                                    | 73/1230 [32:59<8:38:01, 26.86s/it]
  6%|██▎                                    | 74/1230 [33:29<8:53:56, 27.71s/it]
                                                                                
{'loss': 1.0668, 'grad_norm': 0.40338611602783203, 'learning_rate': 0.00017961646364786233, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 168.43, 'epoch': 0.05}

  6%|██▎                                    | 74/1230 [33:29<8:53:56, 27.71s/it]
  6%|██▍                                    | 75/1230 [33:55<8:44:48, 27.26s/it]
                                                                                
{'loss': 1.0724, 'grad_norm': 0.4354206323623657, 'learning_rate': 0.00017959546930174265, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 189.82, 'epoch': 0.05}

  6%|██▍                                    | 75/1230 [33:55<8:44:48, 27.26s/it]
  6%|██▍                                    | 76/1230 [34:20<8:32:48, 26.66s/it]
                                                                                
{'loss': 1.1078, 'grad_norm': 0.4235251545906067, 'learning_rate': 0.00017957391699767323, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 174.1, 'epoch': 0.05}

  6%|██▍                                    | 76/1230 [34:20<8:32:48, 26.66s/it]
  6%|██▍                                    | 77/1230 [34:50<8:47:12, 27.43s/it]
                                                                                
{'loss': 1.0674, 'grad_norm': 0.41283753514289856, 'learning_rate': 0.00017955180688485948, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 163.15, 'epoch': 0.05}

  6%|██▍                                    | 77/1230 [34:50<8:47:12, 27.43s/it]
  6%|██▍                                    | 78/1230 [35:15<8:32:45, 26.71s/it]
                                                                                
{'loss': 1.0594, 'grad_norm': 0.39542660117149353, 'learning_rate': 0.00017952913911636846, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.51, 'epoch': 0.06}

  6%|██▍                                    | 78/1230 [35:15<8:32:45, 26.71s/it]
  6%|██▌                                    | 79/1230 [35:44<8:46:54, 27.47s/it]
                                                                                
{'loss': 1.0922, 'grad_norm': 0.380076140165329, 'learning_rate': 0.0001795059138491279, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 162.91, 'epoch': 0.06}

  6%|██▌                                    | 79/1230 [35:44<8:46:54, 27.47s/it]
  7%|██▌                                    | 80/1230 [36:14<9:00:16, 28.19s/it]
                                                                                
{'loss': 1.091, 'grad_norm': 0.4047469198703766, 'learning_rate': 0.00017948213124392506, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 159.8, 'epoch': 0.06}

  7%|██▌                                    | 80/1230 [36:14<9:00:16, 28.19s/it]
  7%|██▌                                    | 81/1230 [36:44<9:08:38, 28.65s/it]
                                                                                
{'loss': 1.0308, 'grad_norm': 0.3982431888580322, 'learning_rate': 0.00017945779146540553, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 155.86, 'epoch': 0.06}

  7%|██▌                                    | 81/1230 [36:44<9:08:38, 28.65s/it]
  7%|██▌                                    | 82/1230 [37:09<8:52:15, 27.82s/it]
                                                                                
{'loss': 1.0486, 'grad_norm': 0.3937995135784149, 'learning_rate': 0.00017943289468207226, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 185.86, 'epoch': 0.06}

  7%|██▌                                    | 82/1230 [37:09<8:52:15, 27.82s/it]
  7%|██▋                                    | 83/1230 [37:35<8:38:41, 27.13s/it]
                                                                                
{'loss': 1.0993, 'grad_norm': 0.3986201584339142, 'learning_rate': 0.00017940744106628432, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 185.84, 'epoch': 0.06}

  7%|██▋                                    | 83/1230 [37:35<8:38:41, 27.13s/it]
  7%|██▋                                    | 84/1230 [38:05<8:53:59, 27.96s/it]
                                                                                
{'loss': 1.0639, 'grad_norm': 0.41653358936309814, 'learning_rate': 0.00017938143079425562, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 151.71, 'epoch': 0.06}

  7%|██▋                                    | 84/1230 [38:05<8:53:59, 27.96s/it]
  7%|██▋                                    | 85/1230 [38:34<9:01:57, 28.40s/it]
                                                                                
{'loss': 0.9798, 'grad_norm': 0.3905864357948303, 'learning_rate': 0.0001793548640460539, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 164.05, 'epoch': 0.06}

  7%|██▋                                    | 85/1230 [38:34<9:01:57, 28.40s/it]
  7%|██▋                                    | 86/1230 [39:04<9:07:54, 28.74s/it]
                                                                                
{'loss': 1.0805, 'grad_norm': 0.41396772861480713, 'learning_rate': 0.00017932774100559923, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 171.48, 'epoch': 0.06}

  7%|██▋                                    | 86/1230 [39:04<9:07:54, 28.74s/it]
  7%|██▊                                    | 87/1230 [39:33<9:11:32, 28.95s/it]
                                                                                
{'loss': 1.0467, 'grad_norm': 0.40054523944854736, 'learning_rate': 0.0001793000618606629, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 167.29, 'epoch': 0.06}

  7%|██▊                                    | 87/1230 [39:33<9:11:32, 28.95s/it]
  7%|██▊                                    | 88/1230 [39:59<8:53:07, 28.01s/it]
                                                                                
{'loss': 1.0614, 'grad_norm': 0.41317734122276306, 'learning_rate': 0.00017927182680286616, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 188.51, 'epoch': 0.06}

  7%|██▊                                    | 88/1230 [39:59<8:53:07, 28.01s/it]
  7%|██▊                                    | 89/1230 [40:29<9:00:44, 28.43s/it]
                                                                                
{'loss': 0.9809, 'grad_norm': 0.3968685269355774, 'learning_rate': 0.0001792430360276787, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 157.72, 'epoch': 0.06}

  7%|██▊                                    | 89/1230 [40:29<9:00:44, 28.43s/it]
  7%|██▊                                    | 90/1230 [40:54<8:43:00, 27.53s/it]
                                                                                
{'loss': 1.0142, 'grad_norm': 0.5100579857826233, 'learning_rate': 0.00017921368973441743, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 172.09, 'epoch': 0.06}

  7%|██▊                                    | 90/1230 [40:54<8:43:00, 27.53s/it]
  7%|██▉                                    | 91/1230 [41:24<8:55:17, 28.20s/it]
                                                                                
{'loss': 1.017, 'grad_norm': 0.39827218651771545, 'learning_rate': 0.00017918378812624514, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 162.94, 'epoch': 0.06}

  7%|██▉                                    | 91/1230 [41:24<8:55:17, 28.20s/it]
  7%|██▉                                    | 92/1230 [41:50<8:42:03, 27.52s/it]
                                                                                
{'loss': 1.0405, 'grad_norm': 0.42154788970947266, 'learning_rate': 0.00017915333141016899, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 190.64, 'epoch': 0.07}

  7%|██▉                                    | 92/1230 [41:50<8:42:03, 27.52s/it]
  8%|██▉                                    | 93/1230 [42:19<8:52:01, 28.07s/it]
                                                                                
{'loss': 1.0479, 'grad_norm': 0.4218939244747162, 'learning_rate': 0.00017912231979703912, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 159.64, 'epoch': 0.07}

  8%|██▉                                    | 93/1230 [42:19<8:52:01, 28.07s/it]
  8%|██▉                                    | 94/1230 [42:48<8:59:04, 28.47s/it]
                                                                                
{'loss': 1.0093, 'grad_norm': 0.4115059971809387, 'learning_rate': 0.00017909075350154723, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 158.05, 'epoch': 0.07}

  8%|██▉                                    | 94/1230 [42:48<8:59:04, 28.47s/it]
  8%|███                                    | 95/1230 [43:14<8:42:56, 27.64s/it]
                                                                                
{'loss': 1.0492, 'grad_norm': 0.44353851675987244, 'learning_rate': 0.000179058632742225, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 199.13, 'epoch': 0.07}

  8%|███                                    | 95/1230 [43:14<8:42:56, 27.64s/it]
  8%|███                                    | 96/1230 [43:40<8:32:24, 27.11s/it]
                                                                                
{'loss': 1.0248, 'grad_norm': 0.3923754096031189, 'learning_rate': 0.00017902595774144266, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 196.05, 'epoch': 0.07}

  8%|███                                    | 96/1230 [43:40<8:32:24, 27.11s/it]
  8%|███                                    | 97/1230 [44:10<8:46:52, 27.90s/it]
                                                                                
{'loss': 1.016, 'grad_norm': 0.4095909595489502, 'learning_rate': 0.00017899272872540743, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 163.04, 'epoch': 0.07}

  8%|███                                    | 97/1230 [44:10<8:46:52, 27.90s/it]
  8%|███                                    | 98/1230 [44:35<8:31:18, 27.10s/it]
                                                                                
{'loss': 1.037, 'grad_norm': 0.4323609173297882, 'learning_rate': 0.00017895894592416195, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 184.96, 'epoch': 0.07}

  8%|███                                    | 98/1230 [44:35<8:31:18, 27.10s/it]
  8%|███▏                                   | 99/1230 [45:01<8:25:18, 26.81s/it]
                                                                                
{'loss': 1.06, 'grad_norm': 0.4146977961063385, 'learning_rate': 0.00017892460957158267, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 194.53, 'epoch': 0.07}

  8%|███▏                                   | 99/1230 [45:01<8:25:18, 26.81s/it]
  8%|███                                   | 100/1230 [45:26<8:16:58, 26.39s/it]
                                                                                
{'loss': 0.9929, 'grad_norm': 0.41478508710861206, 'learning_rate': 0.00017888971990537823, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 176.65, 'epoch': 0.07}

  8%|███                                   | 100/1230 [45:26<8:16:58, 26.39s/it]
  8%|███                                   | 101/1230 [45:53<8:15:39, 26.34s/it]
                                                                                
{'loss': 1.0879, 'grad_norm': 0.4217905104160309, 'learning_rate': 0.00017885427716708786, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 186.23, 'epoch': 0.07}

  8%|███                                   | 101/1230 [45:53<8:15:39, 26.34s/it]
  8%|███▏                                  | 102/1230 [46:22<8:33:14, 27.30s/it]
                                                                                
{'loss': 0.9776, 'grad_norm': 0.3845559060573578, 'learning_rate': 0.00017881828160207968, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.18, 'tokens_per_second_per_gpu': 168.2, 'epoch': 0.07}

  8%|███▏                                  | 102/1230 [46:22<8:33:14, 27.30s/it]
  8%|███▏                                  | 103/1230 [46:51<8:43:35, 27.88s/it]
                                                                                
{'loss': 1.0062, 'grad_norm': 0.4062027335166931, 'learning_rate': 0.00017878173345954896, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.91, 'epoch': 0.07}

  8%|███▏                                  | 103/1230 [46:51<8:43:35, 27.88s/it]
  8%|███▏                                  | 104/1230 [47:17<8:32:08, 27.29s/it]
                                                                                
{'loss': 1.0095, 'grad_norm': 0.4234515130519867, 'learning_rate': 0.00017874463299251647, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.81, 'epoch': 0.07}

  8%|███▏                                  | 104/1230 [47:17<8:32:08, 27.29s/it]
  9%|███▏                                  | 105/1230 [47:46<8:41:46, 27.83s/it]
                                                                                
{'loss': 0.9876, 'grad_norm': 0.4265342354774475, 'learning_rate': 0.00017870698045782668, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.88, 'epoch': 0.07}

  9%|███▏                                  | 105/1230 [47:46<8:41:46, 27.83s/it]
  9%|███▎                                  | 106/1230 [48:15<8:47:53, 28.18s/it]
                                                                                
{'loss': 1.0396, 'grad_norm': 0.4033720791339874, 'learning_rate': 0.00017866877611614598, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.82, 'epoch': 0.08}

  9%|███▎                                  | 106/1230 [48:15<8:47:53, 28.18s/it]
  9%|███▎                                  | 107/1230 [48:41<8:33:40, 27.44s/it]
                                                                                
{'loss': 1.0234, 'grad_norm': 0.4090272784233093, 'learning_rate': 0.00017863002023196093, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.76, 'epoch': 0.08}

  9%|███▎                                  | 107/1230 [48:41<8:33:40, 27.44s/it]
  9%|███▎                                  | 108/1230 [49:10<8:43:34, 28.00s/it]
                                                                                
{'loss': 1.0472, 'grad_norm': 0.404143363237381, 'learning_rate': 0.0001785907130735763, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.22, 'epoch': 0.08}

  9%|███▎                                  | 108/1230 [49:11<8:43:34, 28.00s/it]
  9%|███▎                                  | 109/1230 [49:36<8:28:20, 27.21s/it]
                                                                                
{'loss': 1.0012, 'grad_norm': 0.4370860457420349, 'learning_rate': 0.00017855085491311342, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.51, 'epoch': 0.08}

  9%|███▎                                  | 109/1230 [49:36<8:28:20, 27.21s/it]
  9%|███▍                                  | 110/1230 [50:06<8:41:43, 27.95s/it]
                                                                                
{'loss': 1.0009, 'grad_norm': 0.41381436586380005, 'learning_rate': 0.0001785104460265081, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.1, 'epoch': 0.08}

  9%|███▍                                  | 110/1230 [50:06<8:41:43, 27.95s/it]
  9%|███▍                                  | 111/1230 [50:32<8:30:39, 27.38s/it]
                                                                                
{'loss': 0.9991, 'grad_norm': 0.4121047556400299, 'learning_rate': 0.00017846948669350882, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.43, 'epoch': 0.08}

  9%|███▍                                  | 111/1230 [50:32<8:30:39, 27.38s/it]
  9%|███▍                                  | 112/1230 [50:58<8:22:36, 26.97s/it]
                                                                                
{'loss': 1.0127, 'grad_norm': 0.4320911765098572, 'learning_rate': 0.00017842797719767474, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.83, 'epoch': 0.08}

  9%|███▍                                  | 112/1230 [50:58<8:22:36, 26.97s/it]
  9%|███▍                                  | 113/1230 [51:27<8:35:41, 27.70s/it]
                                                                                
{'loss': 1.0092, 'grad_norm': 0.4291061758995056, 'learning_rate': 0.00017838591782637383, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.89, 'epoch': 0.08}

  9%|███▍                                  | 113/1230 [51:27<8:35:41, 27.70s/it]
  9%|███▌                                  | 114/1230 [51:53<8:25:11, 27.16s/it]
                                                                                
{'loss': 0.9883, 'grad_norm': 0.43331411480903625, 'learning_rate': 0.00017834330887078073, 'memory/max_active (GiB)': 10.38, 'memory/max_allocated (GiB)': 10.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.89, 'epoch': 0.08}

  9%|███▌                                  | 114/1230 [51:53<8:25:11, 27.16s/it]
  9%|███▌                                  | 115/1230 [52:19<8:18:37, 26.83s/it]
                                                                                
{'loss': 0.9719, 'grad_norm': 0.46419113874435425, 'learning_rate': 0.0001783001506258749, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.25, 'epoch': 0.08}

  9%|███▌                                  | 115/1230 [52:19<8:18:37, 26.83s/it]
  9%|███▌                                  | 116/1230 [52:45<8:16:04, 26.72s/it]
                                                                                
{'loss': 0.9589, 'grad_norm': 0.4231116473674774, 'learning_rate': 0.00017825644339043844, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.29, 'epoch': 0.08}

  9%|███▌                                  | 116/1230 [52:45<8:16:04, 26.72s/it]
 10%|███▌                                  | 117/1230 [53:15<8:28:43, 27.42s/it]
                                                                                
{'loss': 0.9902, 'grad_norm': 0.45582276582717896, 'learning_rate': 0.00017821218746705414, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.31, 'epoch': 0.08}

 10%|███▌                                  | 117/1230 [53:15<8:28:43, 27.42s/it]
 10%|███▋                                  | 118/1230 [53:43<8:36:03, 27.85s/it]
                                                                                
{'loss': 0.9823, 'grad_norm': 0.3943968117237091, 'learning_rate': 0.0001781673831621033, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.95, 'epoch': 0.08}

 10%|███▋                                  | 118/1230 [53:43<8:36:03, 27.85s/it]
 10%|███▋                                  | 119/1230 [54:10<8:26:25, 27.35s/it]
                                                                                
{'loss': 0.9936, 'grad_norm': 0.431410014629364, 'learning_rate': 0.0001781220307857636, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.84, 'epoch': 0.08}

 10%|███▋                                  | 119/1230 [54:10<8:26:25, 27.35s/it]
 10%|███▋                                  | 120/1230 [54:39<8:38:52, 28.05s/it]
                                                                                
{'loss': 0.9611, 'grad_norm': 0.39643439650535583, 'learning_rate': 0.00017807613065200703, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.21, 'epoch': 0.08}

 10%|███▋                                  | 120/1230 [54:39<8:38:52, 28.05s/it]
 10%|███▋                                  | 121/1230 [55:05<8:27:32, 27.46s/it]
                                                                                
{'loss': 0.9553, 'grad_norm': 0.43647265434265137, 'learning_rate': 0.00017802968307859769, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.95, 'epoch': 0.09}

 10%|███▋                                  | 121/1230 [55:05<8:27:32, 27.46s/it]
 10%|███▊                                  | 122/1230 [55:31<8:19:19, 27.04s/it]
                                                                                
{'loss': 0.9578, 'grad_norm': 0.4368506669998169, 'learning_rate': 0.0001779826883870895, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.73, 'epoch': 0.09}

 10%|███▊                                  | 122/1230 [55:31<8:19:19, 27.04s/it]
 10%|███▊                                  | 123/1230 [56:01<8:32:09, 27.76s/it]
                                                                                
{'loss': 0.9606, 'grad_norm': 0.42910128831863403, 'learning_rate': 0.0001779351469028241, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.04, 'epoch': 0.09}

 10%|███▊                                  | 123/1230 [56:01<8:32:09, 27.76s/it]
 10%|███▊                                  | 124/1230 [56:26<8:16:46, 26.95s/it]
                                                                                
{'loss': 0.9812, 'grad_norm': 0.4254802465438843, 'learning_rate': 0.00017788705895492856, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.35, 'epoch': 0.09}

 10%|███▊                                  | 124/1230 [56:26<8:16:46, 26.95s/it]
 10%|███▊                                  | 125/1230 [56:55<8:28:46, 27.63s/it]
                                                                                
{'loss': 0.9347, 'grad_norm': 0.3915286362171173, 'learning_rate': 0.0001778384248763131, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.24, 'epoch': 0.09}

 10%|███▊                                  | 125/1230 [56:55<8:28:46, 27.63s/it]
 10%|███▉                                  | 126/1230 [57:21<8:20:15, 27.19s/it]
                                                                                
{'loss': 1.0316, 'grad_norm': 0.4238954782485962, 'learning_rate': 0.00017778924500366867, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.39, 'epoch': 0.09}

 10%|███▉                                  | 126/1230 [57:21<8:20:15, 27.19s/it]
 10%|███▉                                  | 127/1230 [57:46<8:07:03, 26.49s/it]
                                                                                
{'loss': 0.989, 'grad_norm': 0.44531941413879395, 'learning_rate': 0.0001777395196774648, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.44, 'epoch': 0.09}

 10%|███▉                                  | 127/1230 [57:46<8:07:03, 26.49s/it]
 10%|███▉                                  | 128/1230 [58:15<8:20:00, 27.22s/it]
                                                                                
{'loss': 0.9802, 'grad_norm': 0.42891642451286316, 'learning_rate': 0.00017768924924194726, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.39, 'epoch': 0.09}

 10%|███▉                                  | 128/1230 [58:15<8:20:00, 27.22s/it]
 10%|███▉                                  | 129/1230 [58:41<8:13:45, 26.91s/it]
                                                                                
{'loss': 0.989, 'grad_norm': 0.42402490973472595, 'learning_rate': 0.00017763843404513535, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.05, 'epoch': 0.09}

 10%|███▉                                  | 129/1230 [58:41<8:13:45, 26.91s/it]
 11%|████                                  | 130/1230 [59:10<8:25:49, 27.59s/it]
                                                                                
{'loss': 0.9702, 'grad_norm': 0.43370941281318665, 'learning_rate': 0.0001775870744388199, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.92, 'epoch': 0.09}

 11%|████                                  | 130/1230 [59:10<8:25:49, 27.59s/it]
 11%|████                                  | 131/1230 [59:36<8:12:38, 26.90s/it]
                                                                                
{'loss': 0.9529, 'grad_norm': 0.4282545745372772, 'learning_rate': 0.00017753517077856063, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.34, 'epoch': 0.09}

 11%|████                                  | 131/1230 [59:36<8:12:38, 26.90s/it]
 11%|███▊                                | 132/1230 [1:00:02<8:09:21, 26.74s/it]
                                                                                
{'loss': 0.9466, 'grad_norm': 0.44685453176498413, 'learning_rate': 0.00017748272342368363, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.3, 'epoch': 0.09}

 11%|███▊                                | 132/1230 [1:00:02<8:09:21, 26.74s/it]
 11%|███▉                                | 133/1230 [1:00:32<8:23:57, 27.56s/it]
                                                                                
{'loss': 0.9807, 'grad_norm': 0.42539873719215393, 'learning_rate': 0.00017742973273727908, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.8, 'epoch': 0.09}

 11%|███▉                                | 133/1230 [1:00:32<8:23:57, 27.56s/it]
 11%|███▉                                | 134/1230 [1:00:57<8:10:35, 26.86s/it]
                                                                                
{'loss': 0.9387, 'grad_norm': 0.43209537863731384, 'learning_rate': 0.00017737619908619857, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.24, 'epoch': 0.09}

 11%|███▉                                | 134/1230 [1:00:57<8:10:35, 26.86s/it]
 11%|███▉                                | 135/1230 [1:01:22<8:01:40, 26.39s/it]
                                                                                
{'loss': 0.9213, 'grad_norm': 0.4267615079879761, 'learning_rate': 0.0001773221228410525, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.76, 'epoch': 0.1}

 11%|███▉                                | 135/1230 [1:01:22<8:01:40, 26.39s/it]
 11%|███▉                                | 136/1230 [1:01:52<8:18:07, 27.32s/it]
                                                                                
{'loss': 0.9767, 'grad_norm': 0.4157605767250061, 'learning_rate': 0.00017726750437620776, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.29, 'epoch': 0.1}

 11%|███▉                                | 136/1230 [1:01:52<8:18:07, 27.32s/it]
 11%|████                                | 137/1230 [1:02:21<8:30:59, 28.05s/it]
                                                                                
{'loss': 0.9536, 'grad_norm': 0.41604092717170715, 'learning_rate': 0.00017721234406978498, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.72, 'epoch': 0.1}

 11%|████                                | 137/1230 [1:02:21<8:30:59, 28.05s/it]
 11%|████                                | 138/1230 [1:02:47<8:16:40, 27.29s/it]
                                                                                
{'loss': 0.9326, 'grad_norm': 0.42846816778182983, 'learning_rate': 0.00017715664230365583, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.49, 'epoch': 0.1}

 11%|████                                | 138/1230 [1:02:47<8:16:40, 27.29s/it]
 11%|████                                | 139/1230 [1:03:12<8:05:56, 26.73s/it]
                                                                                
{'loss': 0.9282, 'grad_norm': 0.42142581939697266, 'learning_rate': 0.00017710039946344057, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.9, 'epoch': 0.1}

 11%|████                                | 139/1230 [1:03:12<8:05:56, 26.73s/it]
 11%|████                                | 140/1230 [1:03:37<7:57:36, 26.29s/it]
                                                                                
{'loss': 0.9509, 'grad_norm': 0.434742271900177, 'learning_rate': 0.00017704361593850524, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.24, 'epoch': 0.1}

 11%|████                                | 140/1230 [1:03:37<7:57:36, 26.29s/it]
 11%|████▏                               | 141/1230 [1:04:07<8:16:00, 27.33s/it]
                                                                                
{'loss': 0.9896, 'grad_norm': 0.4472815692424774, 'learning_rate': 0.00017698629212195902, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.8, 'epoch': 0.1}

 11%|████▏                               | 141/1230 [1:04:07<8:16:00, 27.33s/it]
 12%|████▏                               | 142/1230 [1:04:32<8:04:05, 26.70s/it]
                                                                                
{'loss': 0.9409, 'grad_norm': 0.4433630108833313, 'learning_rate': 0.00017692842841065152, 'memory/max_active (GiB)': 12.05, 'memory/max_allocated (GiB)': 12.05, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.49, 'epoch': 0.1}

 12%|████▏                               | 142/1230 [1:04:32<8:04:05, 26.70s/it]
 12%|████▏                               | 143/1230 [1:04:59<8:00:38, 26.53s/it]
                                                                                
{'loss': 0.9523, 'grad_norm': 0.40617841482162476, 'learning_rate': 0.00017687002520517, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.59, 'epoch': 0.1}

 12%|████▏                               | 143/1230 [1:04:59<8:00:38, 26.53s/it]
 12%|████▏                               | 144/1230 [1:05:28<8:17:25, 27.48s/it]
                                                                                
{'loss': 0.9311, 'grad_norm': 0.4168612062931061, 'learning_rate': 0.00017681108290983652, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.74, 'epoch': 0.1}

 12%|████▏                               | 144/1230 [1:05:28<8:17:25, 27.48s/it]
 12%|████▏                               | 145/1230 [1:05:54<8:06:27, 26.90s/it]
                                                                                
{'loss': 0.9552, 'grad_norm': 0.41202038526535034, 'learning_rate': 0.00017675160193270542, 'memory/max_active (GiB)': 7.97, 'memory/max_allocated (GiB)': 7.97, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 209.1, 'epoch': 0.1}

 12%|████▏                               | 145/1230 [1:05:54<8:06:27, 26.90s/it]
 12%|████▎                               | 146/1230 [1:06:19<7:57:58, 26.46s/it]
                                                                                
{'loss': 0.9357, 'grad_norm': 0.42973875999450684, 'learning_rate': 0.00017669158268556015, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.75, 'epoch': 0.1}

 12%|████▎                               | 146/1230 [1:06:19<7:57:58, 26.46s/it]
 12%|████▎                               | 147/1230 [1:06:45<7:54:07, 26.27s/it]
                                                                                
{'loss': 0.9611, 'grad_norm': 0.42357733845710754, 'learning_rate': 0.00017663102558391062, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.8, 'epoch': 0.1}

 12%|████▎                               | 147/1230 [1:06:45<7:54:07, 26.27s/it]
 12%|████▎                               | 148/1230 [1:07:10<7:48:15, 25.97s/it]
                                                                                
{'loss': 0.9457, 'grad_norm': 0.4329891502857208, 'learning_rate': 0.0001765699310469903, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.75, 'epoch': 0.1}

 12%|████▎                               | 148/1230 [1:07:10<7:48:15, 25.97s/it]
 12%|████▎                               | 149/1230 [1:07:40<8:09:02, 27.14s/it]
                                                                                
{'loss': 0.9376, 'grad_norm': 0.4132862687110901, 'learning_rate': 0.00017650829949775326, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.28, 'epoch': 0.11}

 12%|████▎                               | 149/1230 [1:07:40<8:09:02, 27.14s/it]
 12%|████▍                               | 150/1230 [1:08:06<7:58:35, 26.59s/it]
                                                                                
{'loss': 0.9671, 'grad_norm': 0.41959691047668457, 'learning_rate': 0.0001764461313628712, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.5, 'epoch': 0.11}

 12%|████▍                               | 150/1230 [1:08:06<7:58:35, 26.59s/it]
 12%|████▍                               | 151/1230 [1:08:31<7:51:38, 26.23s/it]
                                                                                
{'loss': 0.9085, 'grad_norm': 0.42223286628723145, 'learning_rate': 0.00017638342707273079, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.01, 'epoch': 0.11}

 12%|████▍                               | 151/1230 [1:08:31<7:51:38, 26.23s/it]
 12%|████▍                               | 152/1230 [1:09:00<8:07:57, 27.16s/it]
                                                                                
{'loss': 0.962, 'grad_norm': 0.4067057967185974, 'learning_rate': 0.00017632018706143028, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.9, 'epoch': 0.11}

 12%|████▍                               | 152/1230 [1:09:00<8:07:57, 27.16s/it]
 12%|████▍                               | 153/1230 [1:09:25<7:56:29, 26.55s/it]
                                                                                
{'loss': 0.9423, 'grad_norm': 0.4245084524154663, 'learning_rate': 0.00017625641176677672, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.62, 'epoch': 0.11}

 12%|████▍                               | 153/1230 [1:09:25<7:56:29, 26.55s/it]
 13%|████▌                               | 154/1230 [1:09:51<7:51:21, 26.28s/it]
                                                                                
{'loss': 0.9618, 'grad_norm': 0.41207075119018555, 'learning_rate': 0.000176192101630283, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.12, 'epoch': 0.11}

 13%|████▌                               | 154/1230 [1:09:51<7:51:21, 26.28s/it]
 13%|████▌                               | 155/1230 [1:10:16<7:44:03, 25.90s/it]
                                                                                
{'loss': 0.9386, 'grad_norm': 0.4135086238384247, 'learning_rate': 0.00017612725709716454, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.88, 'epoch': 0.11}

 13%|████▌                               | 155/1230 [1:10:16<7:44:03, 25.90s/it]
 13%|████▌                               | 156/1230 [1:10:42<7:44:00, 25.92s/it]
                                                                                
{'loss': 0.9546, 'grad_norm': 0.41603660583496094, 'learning_rate': 0.0001760618786163365, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.49, 'epoch': 0.11}

 13%|████▌                               | 156/1230 [1:10:42<7:44:00, 25.92s/it]
 13%|████▌                               | 157/1230 [1:11:11<8:00:53, 26.89s/it]
                                                                                
{'loss': 0.9118, 'grad_norm': 0.3927629590034485, 'learning_rate': 0.00017599596664041043, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.49, 'epoch': 0.11}

 13%|████▌                               | 157/1230 [1:11:11<8:00:53, 26.89s/it]
 13%|████▌                               | 158/1230 [1:11:38<7:58:13, 26.77s/it]
                                                                                
{'loss': 0.9263, 'grad_norm': 0.4154106080532074, 'learning_rate': 0.00017592952162569125, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.03, 'epoch': 0.11}

 13%|████▌                               | 158/1230 [1:11:38<7:58:13, 26.77s/it]
 13%|████▋                               | 159/1230 [1:12:07<8:12:41, 27.60s/it]
                                                                                
{'loss': 0.9381, 'grad_norm': 0.41837993264198303, 'learning_rate': 0.00017586254403217417, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.99, 'epoch': 0.11}

 13%|████▋                               | 159/1230 [1:12:07<8:12:41, 27.60s/it]
 13%|████▋                               | 160/1230 [1:12:36<8:19:31, 28.01s/it]
                                                                                
{'loss': 0.9165, 'grad_norm': 0.4128764569759369, 'learning_rate': 0.00017579503432354126, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.48, 'epoch': 0.11}

 13%|████▋                               | 160/1230 [1:12:36<8:19:31, 28.01s/it]
 13%|████▋                               | 161/1230 [1:13:05<8:24:48, 28.33s/it]
                                                                                
{'loss': 0.9364, 'grad_norm': 0.40338653326034546, 'learning_rate': 0.00017572699296715858, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.1, 'epoch': 0.11}

 13%|████▋                               | 161/1230 [1:13:05<8:24:48, 28.33s/it]
 13%|████▋                               | 162/1230 [1:13:35<8:29:58, 28.65s/it]
                                                                                
{'loss': 0.9217, 'grad_norm': 0.40373003482818604, 'learning_rate': 0.00017565842043407254, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.85, 'epoch': 0.11}

 13%|████▋                               | 162/1230 [1:13:35<8:29:58, 28.65s/it]
 13%|████▊                               | 163/1230 [1:14:01<8:15:52, 27.88s/it]
                                                                                
{'loss': 0.9485, 'grad_norm': 0.42709553241729736, 'learning_rate': 0.00017558931719900703, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.92, 'epoch': 0.12}

 13%|████▊                               | 163/1230 [1:14:01<8:15:52, 27.88s/it]
 13%|████▊                               | 164/1230 [1:14:27<8:06:34, 27.39s/it]
                                                                                
{'loss': 0.9206, 'grad_norm': 0.42091599106788635, 'learning_rate': 0.00017551968374035991, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.36, 'epoch': 0.12}

 13%|████▊                               | 164/1230 [1:14:27<8:06:34, 27.39s/it]
 13%|████▊                               | 165/1230 [1:14:53<7:59:41, 27.02s/it]
                                                                                
{'loss': 0.93, 'grad_norm': 0.4177738130092621, 'learning_rate': 0.00017544952054019972, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.17, 'epoch': 0.12}

 13%|████▊                               | 165/1230 [1:14:53<7:59:41, 27.02s/it]
 13%|████▊                               | 166/1230 [1:15:23<8:13:48, 27.85s/it]
                                                                                
{'loss': 0.9443, 'grad_norm': 0.44480565190315247, 'learning_rate': 0.00017537882808426242, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.71, 'epoch': 0.12}

 13%|████▊                               | 166/1230 [1:15:23<8:13:48, 27.85s/it]
 14%|████▉                               | 167/1230 [1:15:53<8:23:08, 28.40s/it]
                                                                                
{'loss': 0.9277, 'grad_norm': 0.4518958032131195, 'learning_rate': 0.0001753076068619479, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.4, 'epoch': 0.12}

 14%|████▉                               | 167/1230 [1:15:53<8:23:08, 28.40s/it]
 14%|████▉                               | 168/1230 [1:16:19<8:10:42, 27.72s/it]
                                                                                
{'loss': 0.911, 'grad_norm': 0.43518558144569397, 'learning_rate': 0.00017523585736631676, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.09, 'epoch': 0.12}

 14%|████▉                               | 168/1230 [1:16:19<8:10:42, 27.72s/it]
 14%|████▉                               | 169/1230 [1:16:45<8:02:50, 27.31s/it]
                                                                                
{'loss': 0.9542, 'grad_norm': 0.4133886992931366, 'learning_rate': 0.00017516358009408672, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.8, 'epoch': 0.12}

 14%|████▉                               | 169/1230 [1:16:45<8:02:50, 27.31s/it]
 14%|████▉                               | 170/1230 [1:17:12<7:58:22, 27.08s/it]
                                                                                
{'loss': 0.9369, 'grad_norm': 0.41094887256622314, 'learning_rate': 0.0001750907755456294, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.21, 'epoch': 0.12}

 14%|████▉                               | 170/1230 [1:17:12<7:58:22, 27.08s/it]
 14%|█████                               | 171/1230 [1:17:37<7:48:37, 26.55s/it]
                                                                                
{'loss': 0.8845, 'grad_norm': 0.4327748417854309, 'learning_rate': 0.00017501744422496654, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.42, 'epoch': 0.12}

 14%|█████                               | 171/1230 [1:17:37<7:48:37, 26.55s/it]
 14%|█████                               | 172/1230 [1:18:03<7:46:52, 26.48s/it]
                                                                                
{'loss': 0.9089, 'grad_norm': 0.44593796133995056, 'learning_rate': 0.00017494358663976684, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.16, 'epoch': 0.12}

 14%|█████                               | 172/1230 [1:18:03<7:46:52, 26.48s/it]
 14%|█████                               | 173/1230 [1:18:33<8:02:11, 27.37s/it]
                                                                                
{'loss': 0.901, 'grad_norm': 0.4389609694480896, 'learning_rate': 0.00017486920330134225, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.07, 'epoch': 0.12}

 14%|█████                               | 173/1230 [1:18:33<8:02:11, 27.37s/it]
 14%|█████                               | 174/1230 [1:19:02<8:12:25, 27.98s/it]
                                                                                
{'loss': 0.9398, 'grad_norm': 0.4476505517959595, 'learning_rate': 0.00017479429472464452, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.25, 'epoch': 0.12}

 14%|█████                               | 174/1230 [1:19:02<8:12:25, 27.98s/it]
 14%|█████                               | 175/1230 [1:19:28<8:00:39, 27.34s/it]
                                                                                
{'loss': 0.9185, 'grad_norm': 0.4365333616733551, 'learning_rate': 0.00017471886142826146, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.64, 'epoch': 0.12}

 14%|█████                               | 175/1230 [1:19:28<8:00:39, 27.34s/it]
 14%|█████▏                              | 176/1230 [1:19:58<8:12:58, 28.06s/it]
                                                                                
{'loss': 0.928, 'grad_norm': 0.3912852704524994, 'learning_rate': 0.0001746429039344137, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.26, 'epoch': 0.12}

 14%|█████▏                              | 176/1230 [1:19:58<8:12:58, 28.06s/it]
 14%|█████▏                              | 177/1230 [1:20:27<8:20:50, 28.54s/it]
                                                                                
{'loss': 0.9397, 'grad_norm': 0.42870280146598816, 'learning_rate': 0.00017456642276895065, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.72, 'epoch': 0.13}

 14%|█████▏                              | 177/1230 [1:20:27<8:20:50, 28.54s/it]
 14%|█████▏                              | 178/1230 [1:20:57<8:26:22, 28.88s/it]
                                                                                
{'loss': 0.9733, 'grad_norm': 0.41583359241485596, 'learning_rate': 0.00017448941846134723, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.7, 'epoch': 0.13}

 14%|█████▏                              | 178/1230 [1:20:57<8:26:22, 28.88s/it]
 15%|█████▏                              | 179/1230 [1:21:26<8:29:02, 29.06s/it]
                                                                                
{'loss': 0.8988, 'grad_norm': 0.4109885096549988, 'learning_rate': 0.00017441189154469993, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.12, 'epoch': 0.13}

 15%|█████▏                              | 179/1230 [1:21:27<8:29:02, 29.06s/it]
 15%|█████▎                              | 180/1230 [1:21:56<8:30:42, 29.18s/it]
                                                                                
{'loss': 0.9452, 'grad_norm': 0.39820384979248047, 'learning_rate': 0.00017433384255572334, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.61, 'epoch': 0.13}

 15%|█████▎                              | 180/1230 [1:21:56<8:30:42, 29.18s/it]
 15%|█████▎                              | 181/1230 [1:22:22<8:11:48, 28.13s/it]
                                                                                
{'loss': 0.878, 'grad_norm': 0.4114847779273987, 'learning_rate': 0.00017425527203474616, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.99, 'epoch': 0.13}

 15%|█████▎                              | 181/1230 [1:22:22<8:11:48, 28.13s/it]
 15%|█████▎                              | 182/1230 [1:22:47<7:58:52, 27.42s/it]
                                                                                
{'loss': 0.9441, 'grad_norm': 0.4220472276210785, 'learning_rate': 0.00017417618052570786, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.25, 'epoch': 0.13}

 15%|█████▎                              | 182/1230 [1:22:47<7:58:52, 27.42s/it]
 15%|█████▎                              | 183/1230 [1:23:17<8:08:47, 28.01s/it]
                                                                                
{'loss': 0.9611, 'grad_norm': 0.4188045561313629, 'learning_rate': 0.0001740965685761545, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.79, 'epoch': 0.13}

 15%|█████▎                              | 183/1230 [1:23:17<8:08:47, 28.01s/it]
 15%|█████▍                              | 184/1230 [1:23:43<7:57:01, 27.36s/it]
                                                                                
{'loss': 0.8971, 'grad_norm': 0.4118848145008087, 'learning_rate': 0.00017401643673723514, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.59, 'epoch': 0.13}

 15%|█████▍                              | 184/1230 [1:23:43<7:57:01, 27.36s/it]
 15%|█████▍                              | 185/1230 [1:24:08<7:46:15, 26.77s/it]
                                                                                
{'loss': 0.9462, 'grad_norm': 0.4361923336982727, 'learning_rate': 0.0001739357855636981, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.39, 'epoch': 0.13}

 15%|█████▍                              | 185/1230 [1:24:08<7:46:15, 26.77s/it]
 15%|█████▍                              | 186/1230 [1:24:34<7:42:34, 26.58s/it]
                                                                                
{'loss': 0.888, 'grad_norm': 0.41224363446235657, 'learning_rate': 0.00017385461561388694, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.72, 'epoch': 0.13}

 15%|█████▍                              | 186/1230 [1:24:34<7:42:34, 26.58s/it]
 15%|█████▍                              | 187/1230 [1:25:00<7:40:21, 26.48s/it]
                                                                                
{'loss': 0.8942, 'grad_norm': 0.41990017890930176, 'learning_rate': 0.00017377292744973674, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.68, 'epoch': 0.13}

 15%|█████▍                              | 187/1230 [1:25:00<7:40:21, 26.48s/it]
 15%|█████▌                              | 188/1230 [1:25:30<7:56:45, 27.45s/it]
                                                                                
{'loss': 0.9395, 'grad_norm': 0.3998962938785553, 'learning_rate': 0.00017369072163677006, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.44, 'epoch': 0.13}

 15%|█████▌                              | 188/1230 [1:25:30<7:56:45, 27.45s/it]
 15%|█████▌                              | 189/1230 [1:25:56<7:49:34, 27.06s/it]
                                                                                
{'loss': 0.957, 'grad_norm': 0.4374053180217743, 'learning_rate': 0.0001736079987440932, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.3, 'epoch': 0.13}

 15%|█████▌                              | 189/1230 [1:25:56<7:49:34, 27.06s/it]
 15%|█████▌                              | 190/1230 [1:26:22<7:39:55, 26.53s/it]
                                                                                
{'loss': 0.9008, 'grad_norm': 0.4396786391735077, 'learning_rate': 0.00017352475934439215, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.66, 'epoch': 0.13}

 15%|█████▌                              | 190/1230 [1:26:22<7:39:55, 26.53s/it]
 16%|█████▌                              | 191/1230 [1:26:51<7:53:54, 27.37s/it]
                                                                                
{'loss': 0.901, 'grad_norm': 0.3822035789489746, 'learning_rate': 0.00017344100401392863, 'memory/max_active (GiB)': 9.87, 'memory/max_allocated (GiB)': 9.87, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.76, 'epoch': 0.14}

 16%|█████▌                              | 191/1230 [1:26:51<7:53:54, 27.37s/it]
 16%|█████▌                              | 192/1230 [1:27:17<7:46:03, 26.94s/it]
                                                                                
{'loss': 0.8248, 'grad_norm': 0.4149349629878998, 'learning_rate': 0.00017335673333253613, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.88, 'epoch': 0.14}

 16%|█████▌                              | 192/1230 [1:27:17<7:46:03, 26.94s/it]
 16%|█████▋                              | 193/1230 [1:27:43<7:43:42, 26.83s/it]
                                                                                
{'loss': 0.8503, 'grad_norm': 0.4126383364200592, 'learning_rate': 0.0001732719478836159, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.75, 'epoch': 0.14}

 16%|█████▋                              | 193/1230 [1:27:43<7:43:42, 26.83s/it]
 16%|█████▋                              | 194/1230 [1:28:13<7:55:23, 27.53s/it]
                                                                                
{'loss': 0.9407, 'grad_norm': 0.42516839504241943, 'learning_rate': 0.0001731866482541329, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.27, 'epoch': 0.14}

 16%|█████▋                              | 194/1230 [1:28:13<7:55:23, 27.53s/it]
 16%|█████▋                              | 195/1230 [1:28:37<7:40:50, 26.72s/it]
                                                                                
{'loss': 0.9077, 'grad_norm': 0.43928098678588867, 'learning_rate': 0.00017310083503461167, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.97, 'epoch': 0.14}

 16%|█████▋                              | 195/1230 [1:28:37<7:40:50, 26.72s/it]
 16%|█████▋                              | 196/1230 [1:29:06<7:51:37, 27.37s/it]
                                                                                
{'loss': 0.888, 'grad_norm': 0.3940887153148651, 'learning_rate': 0.0001730145088191324, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.34, 'epoch': 0.14}

 16%|█████▋                              | 196/1230 [1:29:06<7:51:37, 27.37s/it]
 16%|█████▊                              | 197/1230 [1:29:36<8:00:55, 27.93s/it]
                                                                                
{'loss': 0.9367, 'grad_norm': 0.409182608127594, 'learning_rate': 0.00017292767020532664, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.38, 'epoch': 0.14}

 16%|█████▊                              | 197/1230 [1:29:36<8:00:55, 27.93s/it]
 16%|█████▊                              | 198/1230 [1:30:05<8:07:15, 28.33s/it]
                                                                                
{'loss': 0.9357, 'grad_norm': 0.4328208267688751, 'learning_rate': 0.0001728403197943733, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.79, 'epoch': 0.14}

 16%|█████▊                              | 198/1230 [1:30:05<8:07:15, 28.33s/it]
 16%|█████▊                              | 199/1230 [1:30:31<7:55:04, 27.65s/it]
                                                                                
{'loss': 0.9355, 'grad_norm': 0.4113719165325165, 'learning_rate': 0.00017275245819099445, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.32, 'epoch': 0.14}

 16%|█████▊                              | 199/1230 [1:30:31<7:55:04, 27.65s/it]
 16%|█████▊                              | 200/1230 [1:31:00<8:04:21, 28.21s/it]
                                                                                
{'loss': 0.9142, 'grad_norm': 0.41039717197418213, 'learning_rate': 0.00017266408600345102, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.24, 'epoch': 0.14}

 16%|█████▊                              | 200/1230 [1:31:00<8:04:21, 28.21s/it]
 16%|█████▉                              | 201/1230 [1:31:27<7:55:38, 27.73s/it]
                                                                                
{'loss': 0.8824, 'grad_norm': 0.4228972792625427, 'learning_rate': 0.00017257520384353882, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.46, 'epoch': 0.14}

 16%|█████▉                              | 201/1230 [1:31:27<7:55:38, 27.73s/it]
 16%|█████▉                              | 202/1230 [1:31:53<7:46:28, 27.23s/it]
                                                                                
{'loss': 0.8804, 'grad_norm': 0.3945515751838684, 'learning_rate': 0.000172485812326584, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.65, 'epoch': 0.14}

 16%|█████▉                              | 202/1230 [1:31:53<7:46:28, 27.23s/it]
 17%|█████▉                              | 203/1230 [1:32:22<7:57:17, 27.88s/it]
                                                                                
{'loss': 0.928, 'grad_norm': 0.4105132520198822, 'learning_rate': 0.00017239591207143915, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.28, 'epoch': 0.14}

 17%|█████▉                              | 203/1230 [1:32:22<7:57:17, 27.88s/it]
 17%|█████▉                              | 204/1230 [1:32:49<7:48:03, 27.37s/it]
                                                                                
{'loss': 0.9043, 'grad_norm': 0.40285301208496094, 'learning_rate': 0.00017230550370047866, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.96, 'epoch': 0.14}

 17%|█████▉                              | 204/1230 [1:32:49<7:48:03, 27.37s/it]
 17%|██████                              | 205/1230 [1:33:18<7:57:54, 27.97s/it]
                                                                                
{'loss': 0.9274, 'grad_norm': 0.43520253896713257, 'learning_rate': 0.00017221458783959472, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.53, 'epoch': 0.15}

 17%|██████                              | 205/1230 [1:33:18<7:57:54, 27.97s/it]
 17%|██████                              | 206/1230 [1:33:48<8:05:21, 28.44s/it]
                                                                                
{'loss': 0.9173, 'grad_norm': 0.4620605707168579, 'learning_rate': 0.00017212316511819274, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.22, 'epoch': 0.15}

 17%|██████                              | 206/1230 [1:33:48<8:05:21, 28.44s/it]
 17%|██████                              | 207/1230 [1:34:14<7:53:28, 27.77s/it]
                                                                                
{'loss': 0.8982, 'grad_norm': 0.41204264760017395, 'learning_rate': 0.00017203123616918713, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.11, 'epoch': 0.15}

 17%|██████                              | 207/1230 [1:34:14<7:53:28, 27.77s/it]
 17%|██████                              | 208/1230 [1:34:39<7:40:29, 27.03s/it]
                                                                                
{'loss': 0.8838, 'grad_norm': 0.41535916924476624, 'learning_rate': 0.00017193880162899689, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.24, 'epoch': 0.15}

 17%|██████                              | 208/1230 [1:34:39<7:40:29, 27.03s/it]
 17%|██████                              | 209/1230 [1:35:08<7:51:59, 27.74s/it]
                                                                                
{'loss': 0.919, 'grad_norm': 0.44287437200546265, 'learning_rate': 0.0001718458621375412, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.98, 'epoch': 0.15}

 17%|██████                              | 209/1230 [1:35:08<7:51:59, 27.74s/it]
 17%|██████▏                             | 210/1230 [1:35:38<7:59:17, 28.19s/it]
                                                                                
{'loss': 0.9117, 'grad_norm': 0.4173031747341156, 'learning_rate': 0.00017175241833823502, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.53, 'epoch': 0.15}

 17%|██████▏                             | 210/1230 [1:35:38<7:59:17, 28.19s/it]
 17%|██████▏                             | 211/1230 [1:36:07<8:04:42, 28.54s/it]
                                                                                
{'loss': 0.8858, 'grad_norm': 0.4063745141029358, 'learning_rate': 0.00017165847087798457, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.09, 'epoch': 0.15}

 17%|██████▏                             | 211/1230 [1:36:07<8:04:42, 28.54s/it]
 17%|██████▏                             | 212/1230 [1:36:36<8:08:33, 28.79s/it]
                                                                                
{'loss': 0.9358, 'grad_norm': 0.43398165702819824, 'learning_rate': 0.00017156402040718283, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.99, 'epoch': 0.15}

 17%|██████▏                             | 212/1230 [1:36:36<8:08:33, 28.79s/it]
 17%|██████▏                             | 213/1230 [1:37:06<8:12:23, 29.05s/it]
                                                                                
{'loss': 0.8828, 'grad_norm': 0.424584299325943, 'learning_rate': 0.0001714690675797052, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.08, 'epoch': 0.15}

 17%|██████▏                             | 213/1230 [1:37:06<8:12:23, 29.05s/it]
 17%|██████▎                             | 214/1230 [1:37:35<8:12:50, 29.10s/it]
                                                                                
{'loss': 0.9049, 'grad_norm': 0.4594011604785919, 'learning_rate': 0.0001713736130529048, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.73, 'epoch': 0.15}

 17%|██████▎                             | 214/1230 [1:37:35<8:12:50, 29.10s/it]
 17%|██████▎                             | 215/1230 [1:38:01<7:54:59, 28.08s/it]
                                                                                
{'loss': 0.8885, 'grad_norm': 0.39914393424987793, 'learning_rate': 0.00017127765748760807, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.65, 'epoch': 0.15}

 17%|██████▎                             | 215/1230 [1:38:01<7:54:59, 28.08s/it]
 18%|██████▎                             | 216/1230 [1:38:31<8:02:51, 28.57s/it]
                                                                                
{'loss': 0.8836, 'grad_norm': 0.44385144114494324, 'learning_rate': 0.00017118120154810995, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.41, 'epoch': 0.15}

 18%|██████▎                             | 216/1230 [1:38:31<8:02:51, 28.57s/it]
 18%|██████▎                             | 217/1230 [1:39:00<8:05:30, 28.76s/it]
                                                                                
{'loss': 0.8938, 'grad_norm': 0.4404424726963043, 'learning_rate': 0.0001710842459021696, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.18, 'epoch': 0.15}

 18%|██████▎                             | 217/1230 [1:39:00<8:05:30, 28.76s/it]
 18%|██████▍                             | 218/1230 [1:39:29<8:05:11, 28.77s/it]
                                                                                
{'loss': 0.8804, 'grad_norm': 0.4419240653514862, 'learning_rate': 0.0001709867912210056, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.38, 'epoch': 0.15}

 18%|██████▍                             | 218/1230 [1:39:29<8:05:11, 28.77s/it]
 18%|██████▍                             | 219/1230 [1:39:54<7:48:06, 27.78s/it]
                                                                                
{'loss': 0.8923, 'grad_norm': 0.42205312848091125, 'learning_rate': 0.0001708888381792912, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.99, 'epoch': 0.15}

 18%|██████▍                             | 219/1230 [1:39:54<7:48:06, 27.78s/it]
 18%|██████▍                             | 220/1230 [1:40:20<7:37:29, 27.18s/it]
                                                                                
{'loss': 0.8899, 'grad_norm': 0.4747132956981659, 'learning_rate': 0.0001707903874551499, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.69, 'epoch': 0.16}

 18%|██████▍                             | 220/1230 [1:40:20<7:37:29, 27.18s/it]
 18%|██████▍                             | 221/1230 [1:40:45<7:27:58, 26.64s/it]
                                                                                
{'loss': 0.9151, 'grad_norm': 0.4838443398475647, 'learning_rate': 0.00017069143973015062, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.78, 'epoch': 0.16}

 18%|██████▍                             | 221/1230 [1:40:45<7:27:58, 26.64s/it]
 18%|██████▍                             | 222/1230 [1:41:11<7:23:23, 26.39s/it]
                                                                                
{'loss': 0.8736, 'grad_norm': 0.43420687317848206, 'learning_rate': 0.0001705919956893029, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.31, 'epoch': 0.16}

 18%|██████▍                             | 222/1230 [1:41:11<7:23:23, 26.39s/it]
 18%|██████▌                             | 223/1230 [1:41:36<7:17:24, 26.06s/it]
                                                                                
{'loss': 0.893, 'grad_norm': 0.4052150845527649, 'learning_rate': 0.00017049205602105233, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.2, 'epoch': 0.16}

 18%|██████▌                             | 223/1230 [1:41:36<7:17:24, 26.06s/it]
 18%|██████▌                             | 224/1230 [1:42:02<7:14:53, 25.94s/it]
                                                                                
{'loss': 0.8966, 'grad_norm': 0.41652384400367737, 'learning_rate': 0.00017039162141727562, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.99, 'epoch': 0.16}

 18%|██████▌                             | 224/1230 [1:42:02<7:14:53, 25.94s/it]
 18%|██████▌                             | 225/1230 [1:42:32<7:34:15, 27.12s/it]
                                                                                
{'loss': 0.8909, 'grad_norm': 0.41809943318367004, 'learning_rate': 0.00017029069257327604, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.3, 'epoch': 0.16}

 18%|██████▌                             | 225/1230 [1:42:32<7:34:15, 27.12s/it]
 18%|██████▌                             | 226/1230 [1:42:58<7:26:16, 26.67s/it]
                                                                                
{'loss': 0.8848, 'grad_norm': 0.43565812706947327, 'learning_rate': 0.00017018927018777833, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 201.66, 'epoch': 0.16}

 18%|██████▌                             | 226/1230 [1:42:58<7:26:16, 26.67s/it]
 18%|██████▋                             | 227/1230 [1:43:24<7:23:04, 26.51s/it]
                                                                                
{'loss': 0.8827, 'grad_norm': 0.42474499344825745, 'learning_rate': 0.00017008735496292394, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.35, 'epoch': 0.16}

 18%|██████▋                             | 227/1230 [1:43:24<7:23:04, 26.51s/it]
 19%|██████▋                             | 228/1230 [1:43:49<7:14:30, 26.02s/it]
                                                                                
{'loss': 0.9029, 'grad_norm': 0.4323311746120453, 'learning_rate': 0.0001699849476042664, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.51, 'epoch': 0.16}

 19%|██████▋                             | 228/1230 [1:43:49<7:14:30, 26.02s/it]
 19%|██████▋                             | 229/1230 [1:44:18<7:28:54, 26.91s/it]
                                                                                
{'loss': 0.8443, 'grad_norm': 0.38982218503952026, 'learning_rate': 0.0001698820488207661, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.11, 'epoch': 0.16}

 19%|██████▋                             | 229/1230 [1:44:18<7:28:54, 26.91s/it]
 19%|██████▋                             | 230/1230 [1:44:43<7:19:50, 26.39s/it]
                                                                                
{'loss': 0.9042, 'grad_norm': 0.42199698090553284, 'learning_rate': 0.00016977865932478562, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.66, 'epoch': 0.16}

 19%|██████▋                             | 230/1230 [1:44:43<7:19:50, 26.39s/it]
 19%|██████▊                             | 231/1230 [1:45:12<7:34:57, 27.33s/it]
                                                                                
{'loss': 0.9051, 'grad_norm': 0.4409158527851105, 'learning_rate': 0.0001696747798320847, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.44, 'epoch': 0.16}

 19%|██████▊                             | 231/1230 [1:45:12<7:34:57, 27.33s/it]
 19%|██████▊                             | 232/1230 [1:45:38<7:27:32, 26.91s/it]
                                                                                
{'loss': 0.8897, 'grad_norm': 0.41743943095207214, 'learning_rate': 0.00016957041106181525, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.21, 'epoch': 0.16}

 19%|██████▊                             | 232/1230 [1:45:38<7:27:32, 26.91s/it]
 19%|██████▊                             | 233/1230 [1:46:04<7:21:56, 26.60s/it]
                                                                                
{'loss': 0.9414, 'grad_norm': 0.41664138436317444, 'learning_rate': 0.00016946555373651646, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.06, 'epoch': 0.16}

 19%|██████▊                             | 233/1230 [1:46:04<7:21:56, 26.60s/it]
 19%|██████▊                             | 234/1230 [1:46:34<7:35:49, 27.46s/it]
                                                                                
{'loss': 0.8792, 'grad_norm': 0.41391709446907043, 'learning_rate': 0.00016936020858210983, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.47, 'epoch': 0.17}

 19%|██████▊                             | 234/1230 [1:46:34<7:35:49, 27.46s/it]
 19%|██████▉                             | 235/1230 [1:47:03<7:43:50, 27.97s/it]
                                                                                
{'loss': 0.917, 'grad_norm': 0.4006885886192322, 'learning_rate': 0.0001692543763278939, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.62, 'epoch': 0.17}

 19%|██████▉                             | 235/1230 [1:47:03<7:43:50, 27.97s/it]
 19%|██████▉                             | 236/1230 [1:47:28<7:31:46, 27.27s/it]
                                                                                
{'loss': 0.8302, 'grad_norm': 0.424589604139328, 'learning_rate': 0.00016914805770653954, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 201.71, 'epoch': 0.17}

 19%|██████▉                             | 236/1230 [1:47:28<7:31:46, 27.27s/it]
 19%|██████▉                             | 237/1230 [1:47:53<7:20:10, 26.60s/it]
                                                                                
{'loss': 0.8684, 'grad_norm': 0.4228919446468353, 'learning_rate': 0.00016904125345408466, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.99, 'epoch': 0.17}

 19%|██████▉                             | 237/1230 [1:47:53<7:20:10, 26.60s/it]
 19%|██████▉                             | 238/1230 [1:48:22<7:32:07, 27.35s/it]
                                                                                
{'loss': 0.8748, 'grad_norm': 0.4140578806400299, 'learning_rate': 0.0001689339643099291, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.59, 'epoch': 0.17}

 19%|██████▉                             | 238/1230 [1:48:22<7:32:07, 27.35s/it]
 19%|██████▉                             | 239/1230 [1:48:52<7:41:59, 27.97s/it]
                                                                                
{'loss': 0.8654, 'grad_norm': 0.41670000553131104, 'learning_rate': 0.0001688261910168297, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.48, 'epoch': 0.17}

 19%|██████▉                             | 239/1230 [1:48:52<7:41:59, 27.97s/it]
 20%|███████                             | 240/1230 [1:49:21<7:48:42, 28.41s/it]
                                                                                
{'loss': 0.8451, 'grad_norm': 0.4731596112251282, 'learning_rate': 0.00016871793432089493, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.83, 'epoch': 0.17}

 20%|███████                             | 240/1230 [1:49:21<7:48:42, 28.41s/it]
 20%|███████                             | 241/1230 [1:49:47<7:32:53, 27.48s/it]
                                                                                
{'loss': 0.8772, 'grad_norm': 0.43036189675331116, 'learning_rate': 0.00016860919497157988, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.9, 'epoch': 0.17}

 20%|███████                             | 241/1230 [1:49:47<7:32:53, 27.48s/it]
 20%|███████                             | 242/1230 [1:50:16<7:43:44, 28.16s/it]
                                                                                
{'loss': 0.8487, 'grad_norm': 0.43214133381843567, 'learning_rate': 0.00016849997372168102, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.89, 'epoch': 0.17}

 20%|███████                             | 242/1230 [1:50:16<7:43:44, 28.16s/it]
 20%|███████                             | 243/1230 [1:50:43<7:34:03, 27.60s/it]
                                                                                
{'loss': 0.8175, 'grad_norm': 0.42826151847839355, 'learning_rate': 0.00016839027132733102, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.37, 'epoch': 0.17}

 20%|███████                             | 243/1230 [1:50:43<7:34:03, 27.60s/it]
 20%|███████▏                            | 244/1230 [1:51:08<7:22:51, 26.95s/it]
                                                                                
{'loss': 0.8498, 'grad_norm': 0.41185250878334045, 'learning_rate': 0.00016828008854799335, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.06, 'epoch': 0.17}

 20%|███████▏                            | 244/1230 [1:51:08<7:22:51, 26.95s/it]
 20%|███████▏                            | 245/1230 [1:51:34<7:17:43, 26.66s/it]
                                                                                
{'loss': 0.8382, 'grad_norm': 0.40734440088272095, 'learning_rate': 0.0001681694261464574, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.46, 'epoch': 0.17}

 20%|███████▏                            | 245/1230 [1:51:34<7:17:43, 26.66s/it]
 20%|███████▏                            | 246/1230 [1:52:00<7:13:26, 26.43s/it]
                                                                                
{'loss': 0.8843, 'grad_norm': 0.4257027208805084, 'learning_rate': 0.00016805828488883272, 'memory/max_active (GiB)': 7.96, 'memory/max_allocated (GiB)': 7.96, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 207.56, 'epoch': 0.17}

 20%|███████▏                            | 246/1230 [1:52:00<7:13:26, 26.43s/it]
 20%|███████▏                            | 247/1230 [1:52:27<7:13:53, 26.48s/it]
                                                                                
{'loss': 0.8713, 'grad_norm': 0.425163596868515, 'learning_rate': 0.00016794666554454407, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.89, 'epoch': 0.17}

 20%|███████▏                            | 247/1230 [1:52:27<7:13:53, 26.48s/it]
 20%|███████▎                            | 248/1230 [1:52:52<7:05:51, 26.02s/it]
                                                                                
{'loss': 0.8944, 'grad_norm': 0.451101154088974, 'learning_rate': 0.00016783456888632597, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.56, 'epoch': 0.18}

 20%|███████▎                            | 248/1230 [1:52:52<7:05:51, 26.02s/it]
 20%|███████▎                            | 249/1230 [1:53:16<6:59:57, 25.69s/it]
                                                                                
{'loss': 0.8484, 'grad_norm': 0.45094022154808044, 'learning_rate': 0.00016772199569021738, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.14, 'epoch': 0.18}

 20%|███████▎                            | 249/1230 [1:53:16<6:59:57, 25.69s/it]
 20%|███████▎                            | 250/1230 [1:53:41<6:55:31, 25.44s/it]
                                                                                
{'loss': 0.8305, 'grad_norm': 0.4050379693508148, 'learning_rate': 0.00016760894673555624, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.21, 'epoch': 0.18}

 20%|███████▎                            | 250/1230 [1:53:41<6:55:31, 25.44s/it]
 20%|███████▎                            | 251/1230 [1:54:06<6:53:53, 25.37s/it]
                                                                                
{'loss': 0.7758, 'grad_norm': 2.113538980484009, 'learning_rate': 0.00016749542280497415, 'memory/max_active (GiB)': 2.47, 'memory/max_allocated (GiB)': 2.47, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 3.14, 'epoch': 0.18}

 20%|███████▎                            | 251/1230 [1:54:07<6:53:53, 25.37s/it]
 20%|███████▍                            | 252/1230 [1:54:36<7:15:44, 26.73s/it]
                                                                                
{'loss': 0.8683, 'grad_norm': 0.4240926504135132, 'learning_rate': 0.00016738142468439098, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.42, 'epoch': 0.18}

 20%|███████▍                            | 252/1230 [1:54:36<7:15:44, 26.73s/it]
 21%|███████▍                            | 253/1230 [1:55:06<7:29:35, 27.61s/it]
                                                                                
{'loss': 0.878, 'grad_norm': 0.4495721459388733, 'learning_rate': 0.0001672669531630094, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 142.02, 'epoch': 0.18}

 21%|███████▍                            | 253/1230 [1:55:06<7:29:35, 27.61s/it]
 21%|███████▍                            | 254/1230 [1:55:36<7:39:36, 28.25s/it]
                                                                                
{'loss': 0.8936, 'grad_norm': 0.4282197952270508, 'learning_rate': 0.0001671520090333094, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.84, 'epoch': 0.18}

 21%|███████▍                            | 254/1230 [1:55:36<7:39:36, 28.25s/it]
 21%|███████▍                            | 255/1230 [1:56:05<7:45:02, 28.62s/it]
                                                                                
{'loss': 0.8483, 'grad_norm': 0.40314173698425293, 'learning_rate': 0.00016703659309104274, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.11, 'epoch': 0.18}

 21%|███████▍                            | 255/1230 [1:56:05<7:45:02, 28.62s/it]
 21%|███████▍                            | 256/1230 [1:56:31<7:28:47, 27.65s/it]
                                                                                
{'loss': 0.8696, 'grad_norm': 0.4375151991844177, 'learning_rate': 0.00016692070613522758, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.5, 'epoch': 0.18}

 21%|███████▍                            | 256/1230 [1:56:31<7:28:47, 27.65s/it]
 21%|███████▌                            | 257/1230 [1:56:57<7:20:02, 27.14s/it]
                                                                                
{'loss': 0.8505, 'grad_norm': 0.4504197835922241, 'learning_rate': 0.00016680434896814287, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.91, 'epoch': 0.18}

 21%|███████▌                            | 257/1230 [1:56:57<7:20:02, 27.14s/it]
 21%|███████▌                            | 258/1230 [1:57:22<7:11:45, 26.65s/it]
                                                                                
{'loss': 0.8476, 'grad_norm': 0.40810900926589966, 'learning_rate': 0.00016668752239532285, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.57, 'epoch': 0.18}

 21%|███████▌                            | 258/1230 [1:57:22<7:11:45, 26.65s/it]
 21%|███████▌                            | 259/1230 [1:57:48<7:06:48, 26.37s/it]
                                                                                
{'loss': 0.8771, 'grad_norm': 0.4214816689491272, 'learning_rate': 0.00016657022722555134, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 205.81, 'epoch': 0.18}

 21%|███████▌                            | 259/1230 [1:57:48<7:06:48, 26.37s/it]
 21%|███████▌                            | 260/1230 [1:58:17<7:20:48, 27.27s/it]
                                                                                
{'loss': 0.8469, 'grad_norm': 0.42028114199638367, 'learning_rate': 0.00016645246427085625, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.89, 'epoch': 0.18}

 21%|███████▌                            | 260/1230 [1:58:17<7:20:48, 27.27s/it]
 21%|███████▋                            | 261/1230 [1:58:43<7:14:25, 26.90s/it]
                                                                                
{'loss': 0.8884, 'grad_norm': 0.41013193130493164, 'learning_rate': 0.000166334234346504, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.66, 'epoch': 0.18}

 21%|███████▋                            | 261/1230 [1:58:43<7:14:25, 26.90s/it]
 21%|███████▋                            | 262/1230 [1:59:12<7:24:42, 27.56s/it]
                                                                                
{'loss': 0.8395, 'grad_norm': 0.4069163203239441, 'learning_rate': 0.0001662155382709938, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.13, 'epoch': 0.19}

 21%|███████▋                            | 262/1230 [1:59:12<7:24:42, 27.56s/it]
 21%|███████▋                            | 263/1230 [1:59:42<7:32:41, 28.09s/it]
                                                                                
{'loss': 0.8718, 'grad_norm': 0.40662866830825806, 'learning_rate': 0.00016609637686605195, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.75, 'epoch': 0.19}

 21%|███████▋                            | 263/1230 [1:59:42<7:32:41, 28.09s/it]
 21%|███████▋                            | 264/1230 [2:00:11<7:38:46, 28.50s/it]
                                                                                
{'loss': 0.8523, 'grad_norm': 0.40236273407936096, 'learning_rate': 0.00016597675095662617, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.57, 'epoch': 0.19}

 21%|███████▋                            | 264/1230 [2:00:11<7:38:46, 28.50s/it]
 22%|███████▊                            | 265/1230 [2:00:37<7:24:17, 27.62s/it]
                                                                                
{'loss': 0.8198, 'grad_norm': 0.42605796456336975, 'learning_rate': 0.0001658566613708801, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.34, 'epoch': 0.19}

 22%|███████▊                            | 265/1230 [2:00:37<7:24:17, 27.62s/it]
 22%|███████▊                            | 266/1230 [2:01:02<7:13:34, 26.99s/it]
                                                                                
{'loss': 0.8462, 'grad_norm': 0.4134211838245392, 'learning_rate': 0.00016573610894018715, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.11, 'epoch': 0.19}

 22%|███████▊                            | 266/1230 [2:01:02<7:13:34, 26.99s/it]
 22%|███████▊                            | 267/1230 [2:01:28<7:07:28, 26.63s/it]
                                                                                
{'loss': 0.8703, 'grad_norm': 0.4389660060405731, 'learning_rate': 0.0001656150944991251, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.44, 'epoch': 0.19}

 22%|███████▊                            | 267/1230 [2:01:28<7:07:28, 26.63s/it]
 22%|███████▊                            | 268/1230 [2:01:53<7:00:37, 26.23s/it]
                                                                                
{'loss': 0.8245, 'grad_norm': 0.4312373995780945, 'learning_rate': 0.00016549361888547028, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.03, 'epoch': 0.19}

 22%|███████▊                            | 268/1230 [2:01:53<7:00:37, 26.23s/it]
 22%|███████▊                            | 269/1230 [2:02:18<6:54:34, 25.88s/it]
                                                                                
{'loss': 0.8571, 'grad_norm': 0.4147703945636749, 'learning_rate': 0.00016537168294019153, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.89, 'epoch': 0.19}

 22%|███████▊                            | 269/1230 [2:02:18<6:54:34, 25.88s/it]
 22%|███████▉                            | 270/1230 [2:02:48<7:10:42, 26.92s/it]
                                                                                
{'loss': 0.8849, 'grad_norm': 0.45112377405166626, 'learning_rate': 0.0001652492875074446, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 141.71, 'epoch': 0.19}

 22%|███████▉                            | 270/1230 [2:02:48<7:10:42, 26.92s/it]
 22%|███████▉                            | 271/1230 [2:03:13<7:02:44, 26.45s/it]
                                                                                
{'loss': 0.8743, 'grad_norm': 0.4221040904521942, 'learning_rate': 0.00016512643343456635, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 209.66, 'epoch': 0.19}

 22%|███████▉                            | 271/1230 [2:03:13<7:02:44, 26.45s/it]
 22%|███████▉                            | 272/1230 [2:03:39<6:57:46, 26.17s/it]
                                                                                
{'loss': 0.85, 'grad_norm': 0.3976210653781891, 'learning_rate': 0.00016500312157206857, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 211.97, 'epoch': 0.19}

 22%|███████▉                            | 272/1230 [2:03:39<6:57:46, 26.17s/it]
 22%|███████▉                            | 273/1230 [2:04:04<6:53:09, 25.90s/it]
                                                                                
{'loss': 0.8761, 'grad_norm': 0.43859636783599854, 'learning_rate': 0.00016487935277363256, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.73, 'epoch': 0.19}

 22%|███████▉                            | 273/1230 [2:04:04<6:53:09, 25.90s/it]
 22%|████████                            | 274/1230 [2:04:29<6:51:02, 25.80s/it]
                                                                                
{'loss': 0.8393, 'grad_norm': 0.42936983704566956, 'learning_rate': 0.00016475512789610272, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.08, 'epoch': 0.19}

 22%|████████                            | 274/1230 [2:04:29<6:51:02, 25.80s/it]
 22%|████████                            | 275/1230 [2:04:55<6:50:38, 25.80s/it]
                                                                                
{'loss': 0.8664, 'grad_norm': 0.44082656502723694, 'learning_rate': 0.00016463044779948098, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.73, 'epoch': 0.19}

 22%|████████                            | 275/1230 [2:04:55<6:50:38, 25.80s/it]
 22%|████████                            | 276/1230 [2:05:25<7:08:13, 26.93s/it]
                                                                                
{'loss': 0.8662, 'grad_norm': 0.3813231587409973, 'learning_rate': 0.00016450531334692074, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.51, 'epoch': 0.2}

 22%|████████                            | 276/1230 [2:05:25<7:08:13, 26.93s/it]
 23%|████████                            | 277/1230 [2:05:55<7:20:54, 27.76s/it]
                                                                                
{'loss': 0.8447, 'grad_norm': 0.4445462226867676, 'learning_rate': 0.00016437972540472084, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.2, 'epoch': 0.2}

 23%|████████                            | 277/1230 [2:05:55<7:20:54, 27.76s/it]
 23%|████████▏                           | 278/1230 [2:06:20<7:08:52, 27.03s/it]
                                                                                
{'loss': 0.8567, 'grad_norm': 0.47154346108436584, 'learning_rate': 0.0001642536848423196, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.32, 'epoch': 0.2}

 23%|████████▏                           | 278/1230 [2:06:20<7:08:52, 27.03s/it]
 23%|████████▏                           | 279/1230 [2:06:45<7:01:18, 26.58s/it]
                                                                                
{'loss': 0.8795, 'grad_norm': 0.40929844975471497, 'learning_rate': 0.00016412719253228885, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.34, 'epoch': 0.2}

 23%|████████▏                           | 279/1230 [2:06:45<7:01:18, 26.58s/it]
 23%|████████▏                           | 280/1230 [2:07:15<7:13:49, 27.40s/it]
                                                                                
{'loss': 0.8951, 'grad_norm': 0.42926180362701416, 'learning_rate': 0.00016400024935032782, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.55, 'epoch': 0.2}

 23%|████████▏                           | 280/1230 [2:07:15<7:13:49, 27.40s/it]
 23%|████████▏                           | 281/1230 [2:07:40<7:01:34, 26.65s/it]
                                                                                
{'loss': 0.8599, 'grad_norm': 0.41498661041259766, 'learning_rate': 0.00016387285617525707, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.0, 'epoch': 0.2}

 23%|████████▏                           | 281/1230 [2:07:40<7:01:34, 26.65s/it]
 23%|████████▎                           | 282/1230 [2:08:09<7:14:06, 27.48s/it]
                                                                                
{'loss': 0.894, 'grad_norm': 0.42119100689888, 'learning_rate': 0.00016374501388901247, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.91, 'epoch': 0.2}

 23%|████████▎                           | 282/1230 [2:08:09<7:14:06, 27.48s/it]
 23%|████████▎                           | 283/1230 [2:08:38<7:20:18, 27.90s/it]
                                                                                
{'loss': 0.8669, 'grad_norm': 0.4225495457649231, 'learning_rate': 0.00016361672337663908, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.05, 'epoch': 0.2}

 23%|████████▎                           | 283/1230 [2:08:38<7:20:18, 27.90s/it]
 23%|████████▎                           | 284/1230 [2:09:04<7:10:40, 27.32s/it]
                                                                                
{'loss': 0.8674, 'grad_norm': 0.4159887135028839, 'learning_rate': 0.00016348798552628493, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.72, 'epoch': 0.2}

 23%|████████▎                           | 284/1230 [2:09:04<7:10:40, 27.32s/it]
 23%|████████▎                           | 285/1230 [2:09:34<7:21:41, 28.04s/it]
                                                                                
{'loss': 0.8272, 'grad_norm': 0.41635099053382874, 'learning_rate': 0.00016335880122919506, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.21, 'epoch': 0.2}

 23%|████████▎                           | 285/1230 [2:09:34<7:21:41, 28.04s/it]
 23%|████████▎                           | 286/1230 [2:10:03<7:29:24, 28.56s/it]
                                                                                
{'loss': 0.8051, 'grad_norm': 0.393879771232605, 'learning_rate': 0.00016322917137970508, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.65, 'epoch': 0.2}

 23%|████████▎                           | 286/1230 [2:10:03<7:29:24, 28.56s/it]
 23%|████████▍                           | 287/1230 [2:10:29<7:14:23, 27.64s/it]
                                                                                
{'loss': 0.8824, 'grad_norm': 0.4892124533653259, 'learning_rate': 0.00016309909687523528, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.97, 'epoch': 0.2}

 23%|████████▍                           | 287/1230 [2:10:29<7:14:23, 27.64s/it]
 23%|████████▍                           | 288/1230 [2:10:54<7:03:03, 26.95s/it]
                                                                                
{'loss': 0.8823, 'grad_norm': 0.4207530915737152, 'learning_rate': 0.00016296857861628424, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.88, 'epoch': 0.2}

 23%|████████▍                           | 288/1230 [2:10:54<7:03:03, 26.95s/it]
 23%|████████▍                           | 289/1230 [2:11:24<7:14:52, 27.73s/it]
                                                                                
{'loss': 0.8355, 'grad_norm': 0.4008335471153259, 'learning_rate': 0.00016283761750642258, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.28, 'epoch': 0.2}

 23%|████████▍                           | 289/1230 [2:11:24<7:14:52, 27.73s/it]
 24%|████████▍                           | 290/1230 [2:11:53<7:22:21, 28.24s/it]
                                                                                
{'loss': 0.8653, 'grad_norm': 0.4156545400619507, 'learning_rate': 0.0001627062144522868, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.81, 'epoch': 0.21}

 24%|████████▍                           | 290/1230 [2:11:53<7:22:21, 28.24s/it]
 24%|████████▌                           | 291/1230 [2:12:19<7:10:49, 27.53s/it]
                                                                                
{'loss': 0.8389, 'grad_norm': 0.39179912209510803, 'learning_rate': 0.00016257437036357286, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.17, 'epoch': 0.21}

 24%|████████▌                           | 291/1230 [2:12:19<7:10:49, 27.53s/it]
 24%|████████▌                           | 292/1230 [2:12:45<7:04:25, 27.15s/it]
                                                                                
{'loss': 0.8486, 'grad_norm': 0.39312973618507385, 'learning_rate': 0.00016244208615303013, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.33, 'epoch': 0.21}

 24%|████████▌                           | 292/1230 [2:12:45<7:04:25, 27.15s/it]
 24%|████████▌                           | 293/1230 [2:13:11<6:59:22, 26.85s/it]
                                                                                
{'loss': 0.8546, 'grad_norm': 0.4376573860645294, 'learning_rate': 0.00016230936273645483, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.22, 'epoch': 0.21}

 24%|████████▌                           | 293/1230 [2:13:11<6:59:22, 26.85s/it]
 24%|████████▌                           | 294/1230 [2:13:38<6:55:50, 26.66s/it]
                                                                                
{'loss': 0.8459, 'grad_norm': 0.4089950621128082, 'learning_rate': 0.00016217620103268373, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.22, 'epoch': 0.21}

 24%|████████▌                           | 294/1230 [2:13:38<6:55:50, 26.66s/it]
 24%|████████▋                           | 295/1230 [2:14:07<7:09:15, 27.55s/it]
                                                                                
{'loss': 0.8423, 'grad_norm': 0.3954457640647888, 'learning_rate': 0.00016204260196358794, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.48, 'epoch': 0.21}

 24%|████████▋                           | 295/1230 [2:14:07<7:09:15, 27.55s/it]
 24%|████████▋                           | 296/1230 [2:14:33<6:59:30, 26.95s/it]
                                                                                
{'loss': 0.838, 'grad_norm': 0.41471579670906067, 'learning_rate': 0.00016190856645406637, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.52, 'epoch': 0.21}

 24%|████████▋                           | 296/1230 [2:14:33<6:59:30, 26.95s/it]
 24%|████████▋                           | 297/1230 [2:15:03<7:12:01, 27.78s/it]
                                                                                
{'loss': 0.8654, 'grad_norm': 0.4063762426376343, 'learning_rate': 0.00016177409543203944, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.35, 'epoch': 0.21}

 24%|████████▋                           | 297/1230 [2:15:03<7:12:01, 27.78s/it]
 24%|████████▋                           | 298/1230 [2:15:28<7:00:00, 27.04s/it]
                                                                                
{'loss': 0.8376, 'grad_norm': 0.4331660270690918, 'learning_rate': 0.00016163918982844248, 'memory/max_active (GiB)': 11.54, 'memory/max_allocated (GiB)': 11.54, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.6, 'epoch': 0.21}

 24%|████████▋                           | 298/1230 [2:15:28<7:00:00, 27.04s/it]
 24%|████████▊                           | 299/1230 [2:15:54<6:56:04, 26.81s/it]
                                                                                
{'loss': 0.8013, 'grad_norm': 0.36058327555656433, 'learning_rate': 0.00016150385057721954, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.47, 'epoch': 0.21}

 24%|████████▊                           | 299/1230 [2:15:54<6:56:04, 26.81s/it]
 24%|████████▊                           | 300/1230 [2:16:21<6:53:34, 26.68s/it]
                                                                                
{'loss': 0.8112, 'grad_norm': 0.3999408483505249, 'learning_rate': 0.00016136807861531668, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.63, 'epoch': 0.21}

 24%|████████▊                           | 300/1230 [2:16:21<6:53:34, 26.68s/it]
 24%|████████▊                           | 301/1230 [2:16:49<7:03:24, 27.35s/it]
                                                                                
{'loss': 0.8822, 'grad_norm': 0.4213685393333435, 'learning_rate': 0.00016123187488267563, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.82, 'epoch': 0.21}

 24%|████████▊                           | 301/1230 [2:16:49<7:03:24, 27.35s/it]
 25%|████████▊                           | 302/1230 [2:17:18<7:10:11, 27.81s/it]
                                                                                
{'loss': 0.8356, 'grad_norm': 0.4016689360141754, 'learning_rate': 0.0001610952403222273, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.36, 'epoch': 0.21}

 25%|████████▊                           | 302/1230 [2:17:18<7:10:11, 27.81s/it]
 25%|████████▊                           | 303/1230 [2:17:44<6:57:49, 27.04s/it]
                                                                                
{'loss': 0.8292, 'grad_norm': 0.42766356468200684, 'learning_rate': 0.00016095817587988515, 'memory/max_active (GiB)': 9.76, 'memory/max_allocated (GiB)': 9.76, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.51, 'epoch': 0.21}

 25%|████████▊                           | 303/1230 [2:17:44<6:57:49, 27.04s/it]
 25%|████████▉                           | 304/1230 [2:18:09<6:50:44, 26.61s/it]
                                                                                
{'loss': 0.8219, 'grad_norm': 0.40626150369644165, 'learning_rate': 0.00016082068250453867, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.84, 'epoch': 0.22}

 25%|████████▉                           | 304/1230 [2:18:09<6:50:44, 26.61s/it]
 25%|████████▉                           | 305/1230 [2:18:39<7:04:21, 27.53s/it]
                                                                                
{'loss': 0.8641, 'grad_norm': 0.4103263020515442, 'learning_rate': 0.00016068276114804682, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.19, 'epoch': 0.22}

 25%|████████▉                           | 305/1230 [2:18:39<7:04:21, 27.53s/it]
 25%|████████▉                           | 306/1230 [2:19:04<6:52:31, 26.79s/it]
                                                                                
{'loss': 0.842, 'grad_norm': 0.40893879532814026, 'learning_rate': 0.0001605444127652315, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.58, 'epoch': 0.22}

 25%|████████▉                           | 306/1230 [2:19:04<6:52:31, 26.79s/it]
 25%|████████▉                           | 307/1230 [2:19:30<6:47:04, 26.46s/it]
                                                                                
{'loss': 0.8414, 'grad_norm': 0.422038197517395, 'learning_rate': 0.0001604056383138709, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 201.73, 'epoch': 0.22}

 25%|████████▉                           | 307/1230 [2:19:30<6:47:04, 26.46s/it]
 25%|█████████                           | 308/1230 [2:19:55<6:39:51, 26.02s/it]
                                                                                
{'loss': 0.8393, 'grad_norm': 0.4188705086708069, 'learning_rate': 0.00016026643875469276, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.17, 'epoch': 0.22}

 25%|█████████                           | 308/1230 [2:19:55<6:39:51, 26.02s/it]
 25%|█████████                           | 309/1230 [2:20:21<6:40:30, 26.09s/it]
                                                                                
{'loss': 0.8731, 'grad_norm': 0.39799535274505615, 'learning_rate': 0.00016012681505136794, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.49, 'epoch': 0.22}

 25%|█████████                           | 309/1230 [2:20:21<6:40:30, 26.09s/it]
 25%|█████████                           | 310/1230 [2:20:47<6:38:25, 25.98s/it]
                                                                                
{'loss': 0.8308, 'grad_norm': 0.39051467180252075, 'learning_rate': 0.00015998676817050356, 'memory/max_active (GiB)': 8.61, 'memory/max_allocated (GiB)': 8.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 214.56, 'epoch': 0.22}

 25%|█████████                           | 310/1230 [2:20:47<6:38:25, 25.98s/it]
 25%|█████████                           | 311/1230 [2:21:16<6:53:19, 26.99s/it]
                                                                                
{'loss': 0.8283, 'grad_norm': 0.39024049043655396, 'learning_rate': 0.00015984629908163636, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.39, 'epoch': 0.22}

 25%|█████████                           | 311/1230 [2:21:16<6:53:19, 26.99s/it]
 25%|█████████▏                          | 312/1230 [2:21:45<7:03:53, 27.70s/it]
                                                                                
{'loss': 0.8653, 'grad_norm': 0.4032328426837921, 'learning_rate': 0.00015970540875722601, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.34, 'epoch': 0.22}

 25%|█████████▏                          | 312/1230 [2:21:45<7:03:53, 27.70s/it]
 25%|█████████▏                          | 313/1230 [2:22:14<7:09:58, 28.13s/it]
                                                                                
{'loss': 0.8675, 'grad_norm': 0.4446198344230652, 'learning_rate': 0.00015956409817264846, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.1, 'epoch': 0.22}

 25%|█████████▏                          | 313/1230 [2:22:14<7:09:58, 28.13s/it]
 26%|█████████▏                          | 314/1230 [2:22:44<7:15:29, 28.53s/it]
                                                                                
{'loss': 0.8528, 'grad_norm': 0.4090147614479065, 'learning_rate': 0.00015942236830618895, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.96, 'epoch': 0.22}

 26%|█████████▏                          | 314/1230 [2:22:44<7:15:29, 28.53s/it]
 26%|█████████▏                          | 315/1230 [2:23:09<6:59:48, 27.53s/it]
                                                                                
{'loss': 0.8459, 'grad_norm': 0.39972254633903503, 'learning_rate': 0.0001592802201390355, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.5, 'epoch': 0.22}

 26%|█████████▏                          | 315/1230 [2:23:09<6:59:48, 27.53s/it]
 26%|█████████▏                          | 316/1230 [2:23:35<6:51:29, 27.01s/it]
                                                                                
{'loss': 0.8112, 'grad_norm': 0.4095477759838104, 'learning_rate': 0.00015913765465527195, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.9, 'epoch': 0.22}

 26%|█████████▏                          | 316/1230 [2:23:35<6:51:29, 27.01s/it]
 26%|█████████▎                          | 317/1230 [2:24:01<6:47:10, 26.76s/it]
                                                                                
{'loss': 0.7768, 'grad_norm': 0.3944324254989624, 'learning_rate': 0.0001589946728418712, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.48, 'epoch': 0.22}

 26%|█████████▎                          | 317/1230 [2:24:01<6:47:10, 26.76s/it]
 26%|█████████▎                          | 318/1230 [2:24:31<7:00:02, 27.63s/it]
                                                                                
{'loss': 0.8103, 'grad_norm': 0.41505593061447144, 'learning_rate': 0.00015885127568868843, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.48, 'epoch': 0.23}

 26%|█████████▎                          | 318/1230 [2:24:31<7:00:02, 27.63s/it]
 26%|█████████▎                          | 319/1230 [2:24:56<6:50:22, 27.03s/it]
                                                                                
{'loss': 0.7975, 'grad_norm': 0.40263229608535767, 'learning_rate': 0.0001587074641884541, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.36, 'epoch': 0.23}

 26%|█████████▎                          | 319/1230 [2:24:56<6:50:22, 27.03s/it]
 26%|█████████▎                          | 320/1230 [2:25:23<6:46:19, 26.79s/it]
                                                                                
{'loss': 0.8233, 'grad_norm': 0.44116342067718506, 'learning_rate': 0.00015856323933676728, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.98, 'epoch': 0.23}

 26%|█████████▎                          | 320/1230 [2:25:23<6:46:19, 26.79s/it]
 26%|█████████▍                          | 321/1230 [2:25:49<6:45:02, 26.74s/it]
                                                                                
{'loss': 0.8447, 'grad_norm': 0.4271928369998932, 'learning_rate': 0.00015841860213208853, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.33, 'epoch': 0.23}

 26%|█████████▍                          | 321/1230 [2:25:49<6:45:02, 26.74s/it]
 26%|█████████▍                          | 322/1230 [2:26:14<6:36:55, 26.23s/it]
                                                                                
{'loss': 0.8353, 'grad_norm': 0.41643205285072327, 'learning_rate': 0.00015827355357573322, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.33, 'epoch': 0.23}

 26%|█████████▍                          | 322/1230 [2:26:14<6:36:55, 26.23s/it]
 26%|█████████▍                          | 323/1230 [2:26:40<6:36:25, 26.22s/it]
                                                                                
{'loss': 0.8687, 'grad_norm': 0.4274989366531372, 'learning_rate': 0.00015812809467186442, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.99, 'epoch': 0.23}

 26%|█████████▍                          | 323/1230 [2:26:40<6:36:25, 26.22s/it]
 26%|█████████▍                          | 324/1230 [2:27:05<6:29:34, 25.80s/it]
                                                                                
{'loss': 0.807, 'grad_norm': 0.4006860852241516, 'learning_rate': 0.00015798222642748606, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.67, 'epoch': 0.23}

 26%|█████████▍                          | 324/1230 [2:27:05<6:29:34, 25.80s/it]
 26%|█████████▌                          | 325/1230 [2:27:32<6:31:30, 25.96s/it]
                                                                                
{'loss': 0.8387, 'grad_norm': 0.39527252316474915, 'learning_rate': 0.00015783594985243582, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.94, 'epoch': 0.23}

 26%|█████████▌                          | 325/1230 [2:27:32<6:31:30, 25.96s/it]
 27%|█████████▌                          | 326/1230 [2:28:01<6:45:43, 26.93s/it]
                                                                                
{'loss': 0.8532, 'grad_norm': 0.41433581709861755, 'learning_rate': 0.0001576892659593783, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.57, 'epoch': 0.23}

 27%|█████████▌                          | 326/1230 [2:28:01<6:45:43, 26.93s/it]
 27%|█████████▌                          | 327/1230 [2:28:27<6:43:01, 26.78s/it]
                                                                                
{'loss': 0.7938, 'grad_norm': 0.43573033809661865, 'learning_rate': 0.00015754217576379796, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.07, 'epoch': 0.23}

 27%|█████████▌                          | 327/1230 [2:28:27<6:43:01, 26.78s/it]
 27%|█████████▌                          | 328/1230 [2:28:57<6:55:46, 27.66s/it]
                                                                                
{'loss': 0.8149, 'grad_norm': 0.39664703607559204, 'learning_rate': 0.00015739468028399205, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.2, 'epoch': 0.23}

 27%|█████████▌                          | 328/1230 [2:28:57<6:55:46, 27.66s/it]
 27%|█████████▋                          | 329/1230 [2:29:23<6:49:11, 27.25s/it]
                                                                                
{'loss': 0.8332, 'grad_norm': 0.4115477502346039, 'learning_rate': 0.0001572467805410635, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.86, 'epoch': 0.23}

 27%|█████████▋                          | 329/1230 [2:29:23<6:49:11, 27.25s/it]
 27%|█████████▋                          | 330/1230 [2:29:52<6:58:02, 27.87s/it]
                                                                                
{'loss': 0.8198, 'grad_norm': 0.3887985944747925, 'learning_rate': 0.00015709847755891412, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.25, 'epoch': 0.23}

 27%|█████████▋                          | 330/1230 [2:29:53<6:58:02, 27.87s/it]
 27%|█████████▋                          | 331/1230 [2:30:21<7:01:58, 28.16s/it]
                                                                                
{'loss': 0.824, 'grad_norm': 0.36852923035621643, 'learning_rate': 0.00015694977236423708, 'memory/max_active (GiB)': 7.96, 'memory/max_allocated (GiB)': 7.96, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.78, 'epoch': 0.23}

 27%|█████████▋                          | 331/1230 [2:30:21<7:01:58, 28.16s/it]
 27%|█████████▋                          | 332/1230 [2:30:50<7:05:20, 28.42s/it]
                                                                                
{'loss': 0.8367, 'grad_norm': 0.3903951644897461, 'learning_rate': 0.00015680066598651023, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.32, 'epoch': 0.23}

 27%|█████████▋                          | 332/1230 [2:30:50<7:05:20, 28.42s/it]
 27%|█████████▋                          | 333/1230 [2:31:16<6:54:16, 27.71s/it]
                                                                                
{'loss': 0.7951, 'grad_norm': 0.3898545205593109, 'learning_rate': 0.00015665115945798875, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.91, 'epoch': 0.24}

 27%|█████████▋                          | 333/1230 [2:31:16<6:54:16, 27.71s/it]
 27%|█████████▊                          | 334/1230 [2:31:42<6:44:03, 27.06s/it]
                                                                                
{'loss': 0.8255, 'grad_norm': 0.41010698676109314, 'learning_rate': 0.00015650125381369795, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.97, 'epoch': 0.24}

 27%|█████████▊                          | 334/1230 [2:31:42<6:44:03, 27.06s/it]
 27%|█████████▊                          | 335/1230 [2:32:11<6:53:32, 27.72s/it]
                                                                                
{'loss': 0.8721, 'grad_norm': 0.39171555638313293, 'learning_rate': 0.00015635095009142636, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.12, 'epoch': 0.24}

 27%|█████████▊                          | 335/1230 [2:32:11<6:53:32, 27.72s/it]
 27%|█████████▊                          | 336/1230 [2:32:41<7:00:39, 28.23s/it]
                                                                                
{'loss': 0.8555, 'grad_norm': 0.43433189392089844, 'learning_rate': 0.00015620024933171817, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.52, 'epoch': 0.24}

 27%|█████████▊                          | 336/1230 [2:32:41<7:00:39, 28.23s/it]
 27%|█████████▊                          | 337/1230 [2:33:11<7:08:14, 28.77s/it]
                                                                                
{'loss': 0.8415, 'grad_norm': 0.437129944562912, 'learning_rate': 0.00015604915257786645, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.78, 'epoch': 0.24}

 27%|█████████▊                          | 337/1230 [2:33:11<7:08:14, 28.77s/it]
 27%|█████████▉                          | 338/1230 [2:33:37<6:55:41, 27.96s/it]
                                                                                
{'loss': 0.8562, 'grad_norm': 0.4168522357940674, 'learning_rate': 0.00015589766087590557, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.45, 'epoch': 0.24}

 27%|█████████▉                          | 338/1230 [2:33:37<6:55:41, 27.96s/it]
 28%|█████████▉                          | 339/1230 [2:34:02<6:41:58, 27.07s/it]
                                                                                
{'loss': 0.83, 'grad_norm': 0.42782530188560486, 'learning_rate': 0.00015574577527460417, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.22, 'epoch': 0.24}

 28%|█████████▉                          | 339/1230 [2:34:02<6:41:58, 27.07s/it]
 28%|█████████▉                          | 340/1230 [2:34:31<6:52:24, 27.80s/it]
                                                                                
{'loss': 0.8018, 'grad_norm': 0.4066912829875946, 'learning_rate': 0.00015559349682545785, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.78, 'epoch': 0.24}

 28%|█████████▉                          | 340/1230 [2:34:31<6:52:24, 27.80s/it]
 28%|█████████▉                          | 341/1230 [2:34:57<6:44:49, 27.32s/it]
                                                                                
{'loss': 0.8578, 'grad_norm': 0.40174177289009094, 'learning_rate': 0.00015544082658268177, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.63, 'epoch': 0.24}

 28%|█████████▉                          | 341/1230 [2:34:57<6:44:49, 27.32s/it]
 28%|██████████                          | 342/1230 [2:35:23<6:35:00, 26.69s/it]
                                                                                
{'loss': 0.8155, 'grad_norm': 0.4028593897819519, 'learning_rate': 0.00015528776560320357, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.69, 'epoch': 0.24}

 28%|██████████                          | 342/1230 [2:35:23<6:35:00, 26.69s/it]
 28%|██████████                          | 343/1230 [2:35:49<6:33:25, 26.61s/it]
                                                                                
{'loss': 0.8548, 'grad_norm': 0.4186517596244812, 'learning_rate': 0.00015513431494665588, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.42, 'epoch': 0.24}

 28%|██████████                          | 343/1230 [2:35:49<6:33:25, 26.61s/it]
 28%|██████████                          | 344/1230 [2:36:18<6:44:46, 27.41s/it]
                                                                                
{'loss': 0.8327, 'grad_norm': 0.38873007893562317, 'learning_rate': 0.00015498047567536906, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.65, 'epoch': 0.24}

 28%|██████████                          | 344/1230 [2:36:18<6:44:46, 27.41s/it]
 28%|██████████                          | 345/1230 [2:36:45<6:41:29, 27.22s/it]
                                                                                
{'loss': 0.8055, 'grad_norm': 0.4141141176223755, 'learning_rate': 0.00015482624885436384, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.72, 'epoch': 0.24}

 28%|██████████                          | 345/1230 [2:36:45<6:41:29, 27.22s/it]
 28%|██████████▏                         | 346/1230 [2:37:15<6:50:31, 27.86s/it]
                                                                                
{'loss': 0.8496, 'grad_norm': 0.40166449546813965, 'learning_rate': 0.00015467163555134394, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.18, 'epoch': 0.24}

 28%|██████████▏                         | 346/1230 [2:37:15<6:50:31, 27.86s/it]
 28%|██████████▏                         | 347/1230 [2:37:44<6:56:55, 28.33s/it]
                                                                                
{'loss': 0.7651, 'grad_norm': 0.39091184735298157, 'learning_rate': 0.00015451663683668858, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.87, 'epoch': 0.25}

 28%|██████████▏                         | 347/1230 [2:37:44<6:56:55, 28.33s/it]
 28%|██████████▏                         | 348/1230 [2:38:13<7:01:03, 28.64s/it]
                                                                                
{'loss': 0.7866, 'grad_norm': 0.3928127884864807, 'learning_rate': 0.00015436125378344532, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.39, 'epoch': 0.25}

 28%|██████████▏                         | 348/1230 [2:38:13<7:01:03, 28.64s/it]
 28%|██████████▏                         | 349/1230 [2:38:43<7:04:04, 28.88s/it]
                                                                                
{'loss': 0.8019, 'grad_norm': 0.3984566032886505, 'learning_rate': 0.00015420548746732234, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.08, 'epoch': 0.25}

 28%|██████████▏                         | 349/1230 [2:38:43<7:04:04, 28.88s/it]
 28%|██████████▏                         | 350/1230 [2:39:08<6:46:26, 27.71s/it]
                                                                                
{'loss': 0.8137, 'grad_norm': 0.4161359965801239, 'learning_rate': 0.0001540493389666812, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.87, 'epoch': 0.25}

 28%|██████████▏                         | 350/1230 [2:39:08<6:46:26, 27.71s/it]
 29%|██████████▎                         | 351/1230 [2:39:34<6:39:00, 27.24s/it]
                                                                                
{'loss': 0.8481, 'grad_norm': 0.4320538341999054, 'learning_rate': 0.0001538928093625293, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.37, 'epoch': 0.25}

 29%|██████████▎                         | 351/1230 [2:39:34<6:39:00, 27.24s/it]
 29%|██████████▎                         | 352/1230 [2:39:59<6:29:43, 26.63s/it]
                                                                                
{'loss': 0.804, 'grad_norm': 0.40035656094551086, 'learning_rate': 0.00015373589973851232, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.24, 'epoch': 0.25}

 29%|██████████▎                         | 352/1230 [2:39:59<6:29:43, 26.63s/it]
 29%|██████████▎                         | 353/1230 [2:40:29<6:43:18, 27.59s/it]
                                                                                
{'loss': 0.7619, 'grad_norm': 0.3989311158657074, 'learning_rate': 0.0001535786111809069, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.74, 'epoch': 0.25}

 29%|██████████▎                         | 353/1230 [2:40:29<6:43:18, 27.59s/it]
 29%|██████████▎                         | 354/1230 [2:40:58<6:51:12, 28.16s/it]
                                                                                
{'loss': 0.814, 'grad_norm': 0.40791910886764526, 'learning_rate': 0.00015342094477861296, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.88, 'epoch': 0.25}

 29%|██████████▎                         | 354/1230 [2:40:58<6:51:12, 28.16s/it]
 29%|██████████▍                         | 355/1230 [2:41:24<6:37:48, 27.28s/it]
                                                                                
{'loss': 0.791, 'grad_norm': 0.4250202775001526, 'learning_rate': 0.0001532629016231463, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.61, 'epoch': 0.25}

 29%|██████████▍                         | 355/1230 [2:41:24<6:37:48, 27.28s/it]
 29%|██████████▍                         | 356/1230 [2:41:53<6:46:42, 27.92s/it]
                                                                                
{'loss': 0.8075, 'grad_norm': 0.3936203718185425, 'learning_rate': 0.00015310448280863077, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.22, 'epoch': 0.25}

 29%|██████████▍                         | 356/1230 [2:41:53<6:46:42, 27.92s/it]
 29%|██████████▍                         | 357/1230 [2:42:19<6:38:11, 27.37s/it]
                                                                                
{'loss': 0.8003, 'grad_norm': 0.4207373261451721, 'learning_rate': 0.00015294568943179107, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.05, 'epoch': 0.25}

 29%|██████████▍                         | 357/1230 [2:42:19<6:38:11, 27.37s/it]
 29%|██████████▍                         | 358/1230 [2:42:45<6:29:52, 26.83s/it]
                                                                                
{'loss': 0.8393, 'grad_norm': 0.38205116987228394, 'learning_rate': 0.00015278652259194492, 'memory/max_active (GiB)': 7.96, 'memory/max_allocated (GiB)': 7.96, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 219.73, 'epoch': 0.25}

 29%|██████████▍                         | 358/1230 [2:42:45<6:29:52, 26.83s/it]
 29%|██████████▌                         | 359/1230 [2:43:10<6:23:39, 26.43s/it]
                                                                                
{'loss': 0.8091, 'grad_norm': 0.3968378007411957, 'learning_rate': 0.0001526269833909955, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.42, 'epoch': 0.25}

 29%|██████████▌                         | 359/1230 [2:43:10<6:23:39, 26.43s/it]
 29%|██████████▌                         | 360/1230 [2:43:36<6:19:39, 26.18s/it]
                                                                                
{'loss': 0.7971, 'grad_norm': 0.4055710732936859, 'learning_rate': 0.00015246707293342383, 'memory/max_active (GiB)': 8.86, 'memory/max_allocated (GiB)': 8.86, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 207.8, 'epoch': 0.25}

 29%|██████████▌                         | 360/1230 [2:43:36<6:19:39, 26.18s/it]
 29%|██████████▌                         | 361/1230 [2:44:02<6:17:54, 26.09s/it]
                                                                                
{'loss': 0.8198, 'grad_norm': 0.3915216326713562, 'learning_rate': 0.00015230679232628106, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.89, 'epoch': 0.26}

 29%|██████████▌                         | 361/1230 [2:44:02<6:17:54, 26.09s/it]
 29%|██████████▌                         | 362/1230 [2:44:28<6:18:00, 26.13s/it]
                                                                                
{'loss': 0.7866, 'grad_norm': 0.3921250104904175, 'learning_rate': 0.00015214614267918095, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.41, 'epoch': 0.26}

 29%|██████████▌                         | 362/1230 [2:44:28<6:18:00, 26.13s/it]
 30%|██████████▌                         | 363/1230 [2:44:57<6:29:23, 26.95s/it]
                                                                                
{'loss': 0.8614, 'grad_norm': 0.41333818435668945, 'learning_rate': 0.00015198512510429207, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.46, 'epoch': 0.26}

 30%|██████████▌                         | 363/1230 [2:44:57<6:29:23, 26.95s/it]
 30%|██████████▋                         | 364/1230 [2:45:26<6:37:26, 27.54s/it]
                                                                                
{'loss': 0.8141, 'grad_norm': 0.39629867672920227, 'learning_rate': 0.0001518237407163301, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.09, 'epoch': 0.26}

 30%|██████████▋                         | 364/1230 [2:45:26<6:37:26, 27.54s/it]
 30%|██████████▋                         | 365/1230 [2:45:55<6:44:35, 28.06s/it]
                                                                                
{'loss': 0.7628, 'grad_norm': 0.38778209686279297, 'learning_rate': 0.00015166199063255026, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.32, 'epoch': 0.26}

 30%|██████████▋                         | 365/1230 [2:45:55<6:44:35, 28.06s/it]
 30%|██████████▋                         | 366/1230 [2:46:21<6:34:43, 27.41s/it]
                                                                                
{'loss': 0.8367, 'grad_norm': 0.434656023979187, 'learning_rate': 0.00015149987597273929, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.68, 'epoch': 0.26}

 30%|██████████▋                         | 366/1230 [2:46:21<6:34:43, 27.41s/it]
 30%|██████████▋                         | 367/1230 [2:46:50<6:41:30, 27.92s/it]
                                                                                
{'loss': 0.8289, 'grad_norm': 0.3975108563899994, 'learning_rate': 0.000151337397859208, 'memory/max_active (GiB)': 8.61, 'memory/max_allocated (GiB)': 8.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.85, 'epoch': 0.26}

 30%|██████████▋                         | 367/1230 [2:46:50<6:41:30, 27.92s/it]
 30%|██████████▊                         | 368/1230 [2:47:16<6:33:31, 27.39s/it]
                                                                                
{'loss': 0.7633, 'grad_norm': 0.42445802688598633, 'learning_rate': 0.00015117455741678332, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.57, 'epoch': 0.26}

 30%|██████████▊                         | 368/1230 [2:47:16<6:33:31, 27.39s/it]
 30%|██████████▊                         | 369/1230 [2:47:46<6:42:44, 28.07s/it]
                                                                                
{'loss': 0.8195, 'grad_norm': 0.42212432622909546, 'learning_rate': 0.0001510113557728006, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.29, 'epoch': 0.26}

 30%|██████████▊                         | 369/1230 [2:47:46<6:42:44, 28.07s/it]
 30%|██████████▊                         | 370/1230 [2:48:11<6:29:59, 27.21s/it]
                                                                                
{'loss': 0.8028, 'grad_norm': 0.4296845495700836, 'learning_rate': 0.0001508477940570957, 'memory/max_active (GiB)': 11.54, 'memory/max_allocated (GiB)': 11.54, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.37, 'epoch': 0.26}

 30%|██████████▊                         | 370/1230 [2:48:11<6:29:59, 27.21s/it]
 30%|██████████▊                         | 371/1230 [2:48:40<6:39:32, 27.91s/it]
                                                                                
{'loss': 0.8338, 'grad_norm': 0.3984851837158203, 'learning_rate': 0.0001506838734019973, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.52, 'epoch': 0.26}

 30%|██████████▊                         | 371/1230 [2:48:40<6:39:32, 27.91s/it]
 30%|██████████▉                         | 372/1230 [2:49:10<6:46:00, 28.39s/it]
                                                                                
{'loss': 0.7755, 'grad_norm': 0.39177432656288147, 'learning_rate': 0.00015051959494231896, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.07, 'epoch': 0.26}

 30%|██████████▉                         | 372/1230 [2:49:10<6:46:00, 28.39s/it]
 30%|██████████▉                         | 373/1230 [2:49:35<6:33:07, 27.52s/it]
                                                                                
{'loss': 0.8106, 'grad_norm': 0.40062063932418823, 'learning_rate': 0.00015035495981535133, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 208.58, 'epoch': 0.26}

 30%|██████████▉                         | 373/1230 [2:49:36<6:33:07, 27.52s/it]
 30%|██████████▉                         | 374/1230 [2:50:01<6:23:30, 26.88s/it]
                                                                                
{'loss': 0.8111, 'grad_norm': 0.41415101289749146, 'learning_rate': 0.00015018996916085425, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.22, 'epoch': 0.26}

 30%|██████████▉                         | 374/1230 [2:50:01<6:23:30, 26.88s/it]
 30%|██████████▉                         | 375/1230 [2:50:26<6:16:27, 26.42s/it]
                                                                                
{'loss': 0.8353, 'grad_norm': 0.4233350157737732, 'learning_rate': 0.00015002462412104883, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.43, 'epoch': 0.27}

 30%|██████████▉                         | 375/1230 [2:50:26<6:16:27, 26.42s/it]
 31%|███████████                         | 376/1230 [2:50:52<6:12:00, 26.14s/it]
                                                                                
{'loss': 0.8325, 'grad_norm': 0.4192315340042114, 'learning_rate': 0.00014985892584060956, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.71, 'epoch': 0.27}

 31%|███████████                         | 376/1230 [2:50:52<6:12:00, 26.14s/it]
 31%|███████████                         | 377/1230 [2:51:21<6:27:05, 27.23s/it]
                                                                                
{'loss': 0.8242, 'grad_norm': 0.4104609489440918, 'learning_rate': 0.0001496928754666564, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.09, 'epoch': 0.27}

 31%|███████████                         | 377/1230 [2:51:21<6:27:05, 27.23s/it]
 31%|███████████                         | 378/1230 [2:51:47<6:19:32, 26.73s/it]
                                                                                
{'loss': 0.8064, 'grad_norm': 0.40558961033821106, 'learning_rate': 0.00014952647414874688, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.12, 'epoch': 0.27}

 31%|███████████                         | 378/1230 [2:51:47<6:19:32, 26.73s/it]
 31%|███████████                         | 379/1230 [2:52:12<6:13:26, 26.33s/it]
                                                                                
{'loss': 0.8152, 'grad_norm': 0.42686977982521057, 'learning_rate': 0.00014935972303886812, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.06, 'epoch': 0.27}

 31%|███████████                         | 379/1230 [2:52:12<6:13:26, 26.33s/it]
 31%|███████████                         | 380/1230 [2:52:42<6:26:00, 27.25s/it]
                                                                                
{'loss': 0.7722, 'grad_norm': 0.40077143907546997, 'learning_rate': 0.0001491926232914287, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.9, 'epoch': 0.27}

 31%|███████████                         | 380/1230 [2:52:42<6:26:00, 27.25s/it]
 31%|███████████▏                        | 381/1230 [2:53:08<6:19:54, 26.85s/it]
                                                                                
{'loss': 0.83, 'grad_norm': 0.42426806688308716, 'learning_rate': 0.00014902517606325087, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.49, 'epoch': 0.27}

 31%|███████████▏                        | 381/1230 [2:53:08<6:19:54, 26.85s/it]
 31%|███████████▏                        | 382/1230 [2:53:33<6:13:09, 26.40s/it]
                                                                                
{'loss': 0.7634, 'grad_norm': 0.415157288312912, 'learning_rate': 0.00014885738251356253, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.25, 'epoch': 0.27}

 31%|███████████▏                        | 382/1230 [2:53:33<6:13:09, 26.40s/it]
 31%|███████████▏                        | 383/1230 [2:53:59<6:12:09, 26.36s/it]
                                                                                
{'loss': 0.8093, 'grad_norm': 0.3920951783657074, 'learning_rate': 0.00014868924380398907, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.43, 'epoch': 0.27}

 31%|███████████▏                        | 383/1230 [2:53:59<6:12:09, 26.36s/it]
 31%|███████████▏                        | 384/1230 [2:54:29<6:24:11, 27.25s/it]
                                                                                
{'loss': 0.8125, 'grad_norm': 0.4185640215873718, 'learning_rate': 0.00014852076109854535, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.35, 'epoch': 0.27}

 31%|███████████▏                        | 384/1230 [2:54:29<6:24:11, 27.25s/it]
 31%|███████████▎                        | 385/1230 [2:54:55<6:19:46, 26.97s/it]
                                                                                
{'loss': 0.8332, 'grad_norm': 0.4121361970901489, 'learning_rate': 0.00014835193556362786, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.28, 'epoch': 0.27}

 31%|███████████▎                        | 385/1230 [2:54:55<6:19:46, 26.97s/it]
 31%|███████████▎                        | 386/1230 [2:55:20<6:11:00, 26.37s/it]
                                                                                
{'loss': 0.8241, 'grad_norm': 0.41008374094963074, 'learning_rate': 0.00014818276836800636, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.12, 'epoch': 0.27}

 31%|███████████▎                        | 386/1230 [2:55:20<6:11:00, 26.37s/it]
 31%|███████████▎                        | 387/1230 [2:55:45<6:05:02, 25.98s/it]
                                                                                
{'loss': 0.8422, 'grad_norm': 0.4051925837993622, 'learning_rate': 0.0001480132606828159, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.07, 'epoch': 0.27}

 31%|███████████▎                        | 387/1230 [2:55:45<6:05:02, 25.98s/it]
 32%|███████████▎                        | 388/1230 [2:56:11<6:05:08, 26.02s/it]
                                                                                
{'loss': 0.7863, 'grad_norm': 0.42573726177215576, 'learning_rate': 0.0001478434136815488, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.07, 'epoch': 0.27}

 32%|███████████▎                        | 388/1230 [2:56:11<6:05:08, 26.02s/it]
 32%|███████████▍                        | 389/1230 [2:56:40<6:18:12, 26.98s/it]
                                                                                
{'loss': 0.7969, 'grad_norm': 0.37586864829063416, 'learning_rate': 0.00014767322854004634, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.95, 'epoch': 0.28}

 32%|███████████▍                        | 389/1230 [2:56:40<6:18:12, 26.98s/it]
 32%|███████████▍                        | 390/1230 [2:57:06<6:13:53, 26.71s/it]
                                                                                
{'loss': 0.8122, 'grad_norm': 0.40515390038490295, 'learning_rate': 0.00014750270643649086, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.25, 'epoch': 0.28}

 32%|███████████▍                        | 390/1230 [2:57:06<6:13:53, 26.71s/it]
 32%|███████████▍                        | 391/1230 [2:57:33<6:11:17, 26.55s/it]
                                                                                
{'loss': 0.8172, 'grad_norm': 0.40648603439331055, 'learning_rate': 0.00014733184855139734, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.18, 'epoch': 0.28}

 32%|███████████▍                        | 391/1230 [2:57:33<6:11:17, 26.55s/it]
 32%|███████████▍                        | 392/1230 [2:57:58<6:06:46, 26.26s/it]
                                                                                
{'loss': 0.8732, 'grad_norm': 0.4416288435459137, 'learning_rate': 0.00014716065606760543, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.27, 'epoch': 0.28}

 32%|███████████▍                        | 392/1230 [2:57:58<6:06:46, 26.26s/it]
 32%|███████████▌                        | 393/1230 [2:58:28<6:20:11, 27.25s/it]
                                                                                
{'loss': 0.7978, 'grad_norm': 0.4084922671318054, 'learning_rate': 0.00014698913017027117, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.02, 'epoch': 0.28}

 32%|███████████▌                        | 393/1230 [2:58:28<6:20:11, 27.25s/it]
 32%|███████████▌                        | 394/1230 [2:58:57<6:29:19, 27.94s/it]
                                                                                
{'loss': 0.7654, 'grad_norm': 0.38413211703300476, 'learning_rate': 0.0001468172720468588, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.93, 'epoch': 0.28}

 32%|███████████▌                        | 394/1230 [2:58:57<6:29:19, 27.94s/it]
 32%|███████████▌                        | 395/1230 [2:59:23<6:18:11, 27.18s/it]
                                                                                
{'loss': 0.8424, 'grad_norm': 0.43453723192214966, 'learning_rate': 0.00014664508288713256, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.66, 'epoch': 0.28}

 32%|███████████▌                        | 395/1230 [2:59:23<6:18:11, 27.18s/it]
 32%|███████████▌                        | 396/1230 [2:59:52<6:26:47, 27.83s/it]
                                                                                
{'loss': 0.7867, 'grad_norm': 0.3843488395214081, 'learning_rate': 0.00014647256388314843, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.72, 'epoch': 0.28}

 32%|███████████▌                        | 396/1230 [2:59:52<6:26:47, 27.83s/it]
 32%|███████████▌                        | 397/1230 [3:00:22<6:34:16, 28.40s/it]
                                                                                
{'loss': 0.842, 'grad_norm': 0.4152189791202545, 'learning_rate': 0.00014629971622924582, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.47, 'epoch': 0.28}

 32%|███████████▌                        | 397/1230 [3:00:22<6:34:16, 28.40s/it]
 32%|███████████▋                        | 398/1230 [3:00:47<6:21:34, 27.52s/it]
                                                                                
{'loss': 0.7444, 'grad_norm': 0.37935882806777954, 'learning_rate': 0.00014612654112203946, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 204.98, 'epoch': 0.28}

 32%|███████████▋                        | 398/1230 [3:00:47<6:21:34, 27.52s/it]
 32%|███████████▋                        | 399/1230 [3:01:12<6:11:11, 26.80s/it]
                                                                                
{'loss': 0.8318, 'grad_norm': 0.4415838420391083, 'learning_rate': 0.00014595303976041102, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.04, 'epoch': 0.28}

 32%|███████████▋                        | 399/1230 [3:01:12<6:11:11, 26.80s/it]
 33%|███████████▋                        | 400/1230 [3:01:38<6:07:14, 26.55s/it]
                                                                                
{'loss': 0.8427, 'grad_norm': 0.3987700045108795, 'learning_rate': 0.0001457792133455007, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.63, 'epoch': 0.28}

 33%|███████████▋                        | 400/1230 [3:01:38<6:07:14, 26.55s/it]
 33%|███████████▋                        | 401/1230 [3:02:04<6:02:38, 26.25s/it]
                                                                                
{'loss': 0.7853, 'grad_norm': 0.394652783870697, 'learning_rate': 0.00014560506308069912, 'memory/max_active (GiB)': 8.86, 'memory/max_allocated (GiB)': 8.86, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.82, 'epoch': 0.28}

 33%|███████████▋                        | 401/1230 [3:02:04<6:02:38, 26.25s/it]
 33%|███████████▊                        | 402/1230 [3:02:30<6:02:32, 26.27s/it]
                                                                                
{'loss': 0.8228, 'grad_norm': 0.41577160358428955, 'learning_rate': 0.0001454305901716389, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.36, 'epoch': 0.28}

 33%|███████████▊                        | 402/1230 [3:02:30<6:02:32, 26.27s/it]
 33%|███████████▊                        | 403/1230 [3:03:00<6:15:51, 27.27s/it]
                                                                                
{'loss': 0.7504, 'grad_norm': 0.39295899868011475, 'learning_rate': 0.00014525579582618629, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.73, 'epoch': 0.29}

 33%|███████████▊                        | 403/1230 [3:03:00<6:15:51, 27.27s/it]
 33%|███████████▊                        | 404/1230 [3:03:29<6:24:54, 27.96s/it]
                                                                                
{'loss': 0.7628, 'grad_norm': 0.39601218700408936, 'learning_rate': 0.0001450806812544328, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.85, 'epoch': 0.29}

 33%|███████████▊                        | 404/1230 [3:03:29<6:24:54, 27.96s/it]
 33%|███████████▊                        | 405/1230 [3:03:55<6:13:22, 27.15s/it]
                                                                                
{'loss': 0.7819, 'grad_norm': 0.4208154082298279, 'learning_rate': 0.00014490524766868682, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.5, 'epoch': 0.29}

 33%|███████████▊                        | 405/1230 [3:03:55<6:13:22, 27.15s/it]
 33%|███████████▉                        | 406/1230 [3:04:21<6:08:55, 26.86s/it]
                                                                                
{'loss': 0.799, 'grad_norm': 0.425507515668869, 'learning_rate': 0.00014472949628346535, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.64, 'epoch': 0.29}

 33%|███████████▉                        | 406/1230 [3:04:21<6:08:55, 26.86s/it]
 33%|███████████▉                        | 407/1230 [3:04:46<6:02:27, 26.42s/it]
                                                                                
{'loss': 0.7765, 'grad_norm': 0.4290425181388855, 'learning_rate': 0.0001445534283154855, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.39, 'epoch': 0.29}

 33%|███████████▉                        | 407/1230 [3:04:46<6:02:27, 26.42s/it]
 33%|███████████▉                        | 408/1230 [3:05:13<6:01:17, 26.37s/it]
                                                                                
{'loss': 0.7841, 'grad_norm': 0.40268710255622864, 'learning_rate': 0.00014437704498365594, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.39, 'epoch': 0.29}

 33%|███████████▉                        | 408/1230 [3:05:13<6:01:17, 26.37s/it]
 33%|███████████▉                        | 409/1230 [3:05:42<6:13:31, 27.30s/it]
                                                                                
{'loss': 0.7993, 'grad_norm': 0.3782937228679657, 'learning_rate': 0.00014420034750906865, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.41, 'epoch': 0.29}

 33%|███████████▉                        | 409/1230 [3:05:42<6:13:31, 27.30s/it]
 33%|████████████                        | 410/1230 [3:06:11<6:20:10, 27.82s/it]
                                                                                
{'loss': 0.8317, 'grad_norm': 0.38601264357566833, 'learning_rate': 0.00014402333711499048, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.64, 'epoch': 0.29}

 33%|████████████                        | 410/1230 [3:06:11<6:20:10, 27.82s/it]
 33%|████████████                        | 411/1230 [3:06:37<6:12:07, 27.26s/it]
                                                                                
{'loss': 0.7974, 'grad_norm': 0.39501821994781494, 'learning_rate': 0.00014384601502685453, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.92, 'epoch': 0.29}

 33%|████████████                        | 411/1230 [3:06:37<6:12:07, 27.26s/it]
 33%|████████████                        | 412/1230 [3:07:03<6:07:11, 26.93s/it]
                                                                                
{'loss': 0.7565, 'grad_norm': 0.3856539726257324, 'learning_rate': 0.0001436683824722517, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.55, 'epoch': 0.29}

 33%|████████████                        | 412/1230 [3:07:03<6:07:11, 26.93s/it]
 34%|████████████                        | 413/1230 [3:07:33<6:17:07, 27.70s/it]
                                                                                
{'loss': 0.8132, 'grad_norm': 0.4272928237915039, 'learning_rate': 0.00014349044068092236, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.73, 'epoch': 0.29}

 34%|████████████                        | 413/1230 [3:07:33<6:17:07, 27.70s/it]
 34%|████████████                        | 414/1230 [3:08:02<6:24:03, 28.24s/it]
                                                                                
{'loss': 0.7786, 'grad_norm': 0.41945192217826843, 'learning_rate': 0.00014331219088474753, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 145.34, 'epoch': 0.29}

 34%|████████████                        | 414/1230 [3:08:02<6:24:03, 28.24s/it]
 34%|████████████▏                       | 415/1230 [3:08:32<6:29:30, 28.68s/it]
                                                                                
{'loss': 0.8164, 'grad_norm': 0.4150584042072296, 'learning_rate': 0.00014313363431774073, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.18, 'epoch': 0.29}

 34%|████████████▏                       | 415/1230 [3:08:32<6:29:30, 28.68s/it]
 34%|████████████▏                       | 416/1230 [3:08:58<6:19:23, 27.97s/it]
                                                                                
{'loss': 0.8065, 'grad_norm': 0.4267602264881134, 'learning_rate': 0.00014295477221603907, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.43, 'epoch': 0.29}

 34%|████████████▏                       | 416/1230 [3:08:58<6:19:23, 27.97s/it]
 34%|████████████▏                       | 417/1230 [3:09:27<6:24:32, 28.38s/it]
                                                                                
{'loss': 0.8049, 'grad_norm': 0.39608192443847656, 'learning_rate': 0.00014277560581789494, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.96, 'epoch': 0.3}

 34%|████████████▏                       | 417/1230 [3:09:27<6:24:32, 28.38s/it]
 34%|████████████▏                       | 418/1230 [3:09:57<6:27:56, 28.67s/it]
                                                                                
{'loss': 0.789, 'grad_norm': 0.3989132046699524, 'learning_rate': 0.00014259613636366742, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.2, 'epoch': 0.3}

 34%|████████████▏                       | 418/1230 [3:09:57<6:27:56, 28.67s/it]
 34%|████████████▎                       | 419/1230 [3:10:23<6:16:10, 27.83s/it]
                                                                                
{'loss': 0.7723, 'grad_norm': 0.3921254873275757, 'learning_rate': 0.00014241636509581342, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.8, 'epoch': 0.3}

 34%|████████████▎                       | 419/1230 [3:10:23<6:16:10, 27.83s/it]
 34%|████████████▎                       | 420/1230 [3:10:52<6:23:46, 28.43s/it]
                                                                                
{'loss': 0.7895, 'grad_norm': 0.377063125371933, 'learning_rate': 0.00014223629325887954, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.6, 'epoch': 0.3}

 34%|████████████▎                       | 420/1230 [3:10:53<6:23:46, 28.43s/it]
 34%|████████████▎                       | 421/1230 [3:11:22<6:26:47, 28.69s/it]
                                                                                
{'loss': 0.7884, 'grad_norm': 0.39322835206985474, 'learning_rate': 0.00014205592209949302, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.9, 'epoch': 0.3}

 34%|████████████▎                       | 421/1230 [3:11:22<6:26:47, 28.69s/it]
 34%|████████████▎                       | 422/1230 [3:11:51<6:29:10, 28.90s/it]
                                                                                
{'loss': 0.8473, 'grad_norm': 0.4406552314758301, 'learning_rate': 0.0001418752528663534, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.55, 'epoch': 0.3}

 34%|████████████▎                       | 422/1230 [3:11:51<6:29:10, 28.90s/it]
 34%|████████████▍                       | 423/1230 [3:12:16<6:13:14, 27.75s/it]
                                                                                
{'loss': 0.7818, 'grad_norm': 0.394330769777298, 'learning_rate': 0.00014169428681022375, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.83, 'epoch': 0.3}

 34%|████████████▍                       | 423/1230 [3:12:16<6:13:14, 27.75s/it]
 34%|████████████▍                       | 424/1230 [3:12:42<6:02:59, 27.02s/it]
                                                                                
{'loss': 0.7535, 'grad_norm': 0.4132041931152344, 'learning_rate': 0.00014151302518392194, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.0, 'epoch': 0.3}

 34%|████████████▍                       | 424/1230 [3:12:42<6:02:59, 27.02s/it]
 35%|████████████▍                       | 425/1230 [3:13:07<5:55:28, 26.49s/it]
                                                                                
{'loss': 0.7651, 'grad_norm': 0.41436967253685, 'learning_rate': 0.00014133146924231218, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.87, 'epoch': 0.3}

 35%|████████████▍                       | 425/1230 [3:13:07<5:55:28, 26.49s/it]
 35%|████████████▍                       | 426/1230 [3:13:37<6:08:23, 27.49s/it]
                                                                                
{'loss': 0.7723, 'grad_norm': 0.39960968494415283, 'learning_rate': 0.00014114962024229614, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.21, 'epoch': 0.3}

 35%|████████████▍                       | 426/1230 [3:13:37<6:08:23, 27.49s/it]
 35%|████████████▍                       | 427/1230 [3:14:03<6:01:25, 27.01s/it]
                                                                                
{'loss': 0.7406, 'grad_norm': 0.39408808946609497, 'learning_rate': 0.00014096747944280429, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.52, 'epoch': 0.3}

 35%|████████████▍                       | 427/1230 [3:14:03<6:01:25, 27.01s/it]
 35%|████████████▌                       | 428/1230 [3:14:28<5:54:46, 26.54s/it]
                                                                                
{'loss': 0.7751, 'grad_norm': 0.39944252371788025, 'learning_rate': 0.00014078504810478726, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.56, 'epoch': 0.3}

 35%|████████████▌                       | 428/1230 [3:14:28<5:54:46, 26.54s/it]
 35%|████████████▌                       | 429/1230 [3:14:54<5:51:25, 26.32s/it]
                                                                                
{'loss': 0.7821, 'grad_norm': 0.4035184383392334, 'learning_rate': 0.00014060232749120712, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.44, 'epoch': 0.3}

 35%|████████████▌                       | 429/1230 [3:14:54<5:51:25, 26.32s/it]
 35%|████████████▌                       | 430/1230 [3:15:20<5:51:04, 26.33s/it]
                                                                                
{'loss': 0.7593, 'grad_norm': 0.3973766565322876, 'learning_rate': 0.0001404193188670285, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.59, 'epoch': 0.3}

 35%|████████████▌                       | 430/1230 [3:15:20<5:51:04, 26.33s/it]
 35%|████████████▌                       | 431/1230 [3:15:47<5:51:41, 26.41s/it]
                                                                                
{'loss': 0.7469, 'grad_norm': 0.3945920467376709, 'learning_rate': 0.0001402360234992098, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.84, 'epoch': 0.31}

 35%|████████████▌                       | 431/1230 [3:15:47<5:51:41, 26.41s/it]
 35%|████████████▋                       | 432/1230 [3:16:12<5:46:45, 26.07s/it]
                                                                                
{'loss': 0.763, 'grad_norm': 0.3978945314884186, 'learning_rate': 0.0001400524426566948, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.37, 'epoch': 0.31}

 35%|████████████▋                       | 432/1230 [3:16:12<5:46:45, 26.07s/it]
 35%|████████████▋                       | 433/1230 [3:16:37<5:43:17, 25.84s/it]
                                                                                
{'loss': 0.8016, 'grad_norm': 0.41250884532928467, 'learning_rate': 0.00013986857761040338, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.34, 'epoch': 0.31}

 35%|████████████▋                       | 433/1230 [3:16:37<5:43:17, 25.84s/it]
 35%|████████████▋                       | 434/1230 [3:17:03<5:40:26, 25.66s/it]
                                                                                
{'loss': 0.8292, 'grad_norm': 0.4329974949359894, 'learning_rate': 0.00013968442963322307, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.9, 'epoch': 0.31}

 35%|████████████▋                       | 434/1230 [3:17:03<5:40:26, 25.66s/it]
 35%|████████████▋                       | 435/1230 [3:17:32<5:54:33, 26.76s/it]
                                                                                
{'loss': 0.7475, 'grad_norm': 0.36097198724746704, 'learning_rate': 0.0001395, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.68, 'epoch': 0.31}

 35%|████████████▋                       | 435/1230 [3:17:32<5:54:33, 26.76s/it]
 35%|████████████▊                       | 436/1230 [3:17:58<5:50:17, 26.47s/it]
                                                                                
{'loss': 0.7807, 'grad_norm': 0.3876938223838806, 'learning_rate': 0.00013931528998753034, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 206.39, 'epoch': 0.31}

 35%|████████████▊                       | 436/1230 [3:17:58<5:50:17, 26.47s/it]
 36%|████████████▊                       | 437/1230 [3:18:27<6:01:46, 27.37s/it]
                                                                                
{'loss': 0.8244, 'grad_norm': 0.39826083183288574, 'learning_rate': 0.0001391303008745512, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.43, 'epoch': 0.31}

 36%|████████████▊                       | 437/1230 [3:18:27<6:01:46, 27.37s/it]
 36%|████████████▊                       | 438/1230 [3:18:53<5:54:07, 26.83s/it]
                                                                                
{'loss': 0.8243, 'grad_norm': 0.41090548038482666, 'learning_rate': 0.00013894503394173188, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.5, 'epoch': 0.31}

 36%|████████████▊                       | 438/1230 [3:18:53<5:54:07, 26.83s/it]
 36%|████████████▊                       | 439/1230 [3:19:18<5:47:43, 26.38s/it]
                                                                                
{'loss': 0.8053, 'grad_norm': 0.43880099058151245, 'learning_rate': 0.0001387594904716651, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.26, 'epoch': 0.31}

 36%|████████████▊                       | 439/1230 [3:19:18<5:47:43, 26.38s/it]
 36%|████████████▉                       | 440/1230 [3:19:43<5:43:07, 26.06s/it]
                                                                                
{'loss': 0.7535, 'grad_norm': 0.4031776785850525, 'learning_rate': 0.00013857367174885794, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.58, 'epoch': 0.31}

 36%|████████████▉                       | 440/1230 [3:19:43<5:43:07, 26.06s/it]
 36%|████████████▉                       | 441/1230 [3:20:09<5:40:27, 25.89s/it]
                                                                                
{'loss': 0.7916, 'grad_norm': 0.4165748357772827, 'learning_rate': 0.00013838757905972312, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.32, 'epoch': 0.31}

 36%|████████████▉                       | 441/1230 [3:20:09<5:40:27, 25.89s/it]
 36%|████████████▉                       | 442/1230 [3:20:38<5:53:32, 26.92s/it]
                                                                                
{'loss': 0.7582, 'grad_norm': 0.3813152313232422, 'learning_rate': 0.00013820121369256995, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.56, 'epoch': 0.31}

 36%|████████████▉                       | 442/1230 [3:20:38<5:53:32, 26.92s/it]
 36%|████████████▉                       | 443/1230 [3:21:07<6:02:26, 27.63s/it]
                                                                                
{'loss': 0.7845, 'grad_norm': 0.39812061190605164, 'learning_rate': 0.0001380145769375955, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.17, 'epoch': 0.31}

 36%|████████████▉                       | 443/1230 [3:21:07<6:02:26, 27.63s/it]
 36%|████████████▉                       | 444/1230 [3:21:33<5:51:51, 26.86s/it]
                                                                                
{'loss': 0.7783, 'grad_norm': 0.39913031458854675, 'learning_rate': 0.00013782767008687564, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.26, 'epoch': 0.31}

 36%|████████████▉                       | 444/1230 [3:21:33<5:51:51, 26.86s/it]
 36%|█████████████                       | 445/1230 [3:21:57<5:43:33, 26.26s/it]
                                                                                
{'loss': 0.811, 'grad_norm': 0.42293581366539, 'learning_rate': 0.00013764049443435608, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.1, 'epoch': 0.31}

 36%|█████████████                       | 445/1230 [3:21:57<5:43:33, 26.26s/it]
 36%|█████████████                       | 446/1230 [3:22:26<5:53:54, 27.09s/it]
                                                                                
{'loss': 0.8498, 'grad_norm': 0.41322046518325806, 'learning_rate': 0.00013745305127584346, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.29, 'epoch': 0.32}

 36%|█████████████                       | 446/1230 [3:22:26<5:53:54, 27.09s/it]
 36%|█████████████                       | 447/1230 [3:22:55<6:00:16, 27.61s/it]
                                                                                
{'loss': 0.7671, 'grad_norm': 0.4150811433792114, 'learning_rate': 0.00013726534190899636, 'memory/max_active (GiB)': 12.05, 'memory/max_allocated (GiB)': 12.05, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 146.1, 'epoch': 0.32}

 36%|█████████████                       | 447/1230 [3:22:55<6:00:16, 27.61s/it]
 36%|█████████████                       | 448/1230 [3:23:24<6:05:13, 28.02s/it]
                                                                                
{'loss': 0.7804, 'grad_norm': 0.39436864852905273, 'learning_rate': 0.00013707736763331625, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.08, 'epoch': 0.32}

 36%|█████████████                       | 448/1230 [3:23:24<6:05:13, 28.02s/it]
 37%|█████████████▏                      | 449/1230 [3:23:53<6:09:00, 28.35s/it]
                                                                                
{'loss': 0.8171, 'grad_norm': 0.4019521176815033, 'learning_rate': 0.0001368891297501386, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.24, 'epoch': 0.32}

 37%|█████████████▏                      | 449/1230 [3:23:53<6:09:00, 28.35s/it]
 37%|█████████████▏                      | 450/1230 [3:24:19<5:58:08, 27.55s/it]
                                                                                
{'loss': 0.7523, 'grad_norm': 0.40726417303085327, 'learning_rate': 0.00013670062956262384, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.32, 'epoch': 0.32}

 37%|█████████████▏                      | 450/1230 [3:24:19<5:58:08, 27.55s/it]
 37%|█████████████▏                      | 451/1230 [3:24:45<5:52:19, 27.14s/it]
                                                                                
{'loss': 0.7604, 'grad_norm': 0.3949551284313202, 'learning_rate': 0.00013651186837574817, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.7, 'epoch': 0.32}

 37%|█████████████▏                      | 451/1230 [3:24:45<5:52:19, 27.14s/it]
 37%|█████████████▏                      | 452/1230 [3:25:11<5:47:16, 26.78s/it]
                                                                                
{'loss': 0.8298, 'grad_norm': 0.4061610698699951, 'learning_rate': 0.00013632284749629488, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.55, 'epoch': 0.32}

 37%|█████████████▏                      | 452/1230 [3:25:11<5:47:16, 26.78s/it]
 37%|█████████████▎                      | 453/1230 [3:25:37<5:44:18, 26.59s/it]
                                                                                
{'loss': 0.7448, 'grad_norm': 0.39159470796585083, 'learning_rate': 0.000136133568232845, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.61, 'epoch': 0.32}

 37%|█████████████▎                      | 453/1230 [3:25:37<5:44:18, 26.59s/it]
 37%|█████████████▎                      | 454/1230 [3:26:03<5:41:32, 26.41s/it]
                                                                                
{'loss': 0.7564, 'grad_norm': 0.37709930539131165, 'learning_rate': 0.0001359440318957683, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 203.8, 'epoch': 0.32}

 37%|█████████████▎                      | 454/1230 [3:26:03<5:41:32, 26.41s/it]
 37%|█████████████▎                      | 455/1230 [3:26:28<5:35:41, 25.99s/it]
                                                                                
{'loss': 0.8045, 'grad_norm': 0.40057140588760376, 'learning_rate': 0.00013575423979721429, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.17, 'epoch': 0.32}

 37%|█████████████▎                      | 455/1230 [3:26:28<5:35:41, 25.99s/it]
 37%|█████████████▎                      | 456/1230 [3:26:55<5:37:11, 26.14s/it]
                                                                                
{'loss': 0.7815, 'grad_norm': 0.41337376832962036, 'learning_rate': 0.00013556419325110315, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.18, 'epoch': 0.32}

 37%|█████████████▎                      | 456/1230 [3:26:55<5:37:11, 26.14s/it]
 37%|█████████████▍                      | 457/1230 [3:27:24<5:47:41, 26.99s/it]
                                                                                
{'loss': 0.7885, 'grad_norm': 0.3882520794868469, 'learning_rate': 0.00013537389357311656, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.84, 'epoch': 0.32}

 37%|█████████████▍                      | 457/1230 [3:27:24<5:47:41, 26.99s/it]
 37%|█████████████▍                      | 458/1230 [3:27:49<5:39:28, 26.38s/it]
                                                                                
{'loss': 0.7726, 'grad_norm': 0.3949267864227295, 'learning_rate': 0.00013518334208068852, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.41, 'epoch': 0.32}

 37%|█████████████▍                      | 458/1230 [3:27:49<5:39:28, 26.38s/it]
 37%|█████████████▍                      | 459/1230 [3:28:15<5:38:37, 26.35s/it]
                                                                                
{'loss': 0.7825, 'grad_norm': 0.4016813039779663, 'learning_rate': 0.00013499254009299646, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.55, 'epoch': 0.32}

 37%|█████████████▍                      | 459/1230 [3:28:15<5:38:37, 26.35s/it]
 37%|█████████████▍                      | 460/1230 [3:28:44<5:50:02, 27.28s/it]
                                                                                
{'loss': 0.8042, 'grad_norm': 0.40750497579574585, 'learning_rate': 0.00013480148893095192, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.01, 'epoch': 0.33}

 37%|█████████████▍                      | 460/1230 [3:28:44<5:50:02, 27.28s/it]
 37%|█████████████▍                      | 461/1230 [3:29:10<5:42:18, 26.71s/it]
                                                                                
{'loss': 0.7536, 'grad_norm': 0.4007730185985565, 'learning_rate': 0.00013461018991719147, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.02, 'epoch': 0.33}

 37%|█████████████▍                      | 461/1230 [3:29:10<5:42:18, 26.71s/it]
 38%|█████████████▌                      | 462/1230 [3:29:35<5:37:13, 26.35s/it]
                                                                                
{'loss': 0.7671, 'grad_norm': 0.40316733717918396, 'learning_rate': 0.00013441864437606753, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.15, 'epoch': 0.33}

 38%|█████████████▌                      | 462/1230 [3:29:35<5:37:13, 26.35s/it]
 38%|█████████████▌                      | 463/1230 [3:30:01<5:33:33, 26.09s/it]
                                                                                
{'loss': 0.802, 'grad_norm': 0.4061115086078644, 'learning_rate': 0.00013422685363363924, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.8, 'epoch': 0.33}

 38%|█████████████▌                      | 463/1230 [3:30:01<5:33:33, 26.09s/it]
 38%|█████████████▌                      | 464/1230 [3:30:26<5:30:24, 25.88s/it]
                                                                                
{'loss': 0.7937, 'grad_norm': 0.41946399211883545, 'learning_rate': 0.00013403481901766323, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.21, 'epoch': 0.33}

 38%|█████████████▌                      | 464/1230 [3:30:26<5:30:24, 25.88s/it]
 38%|█████████████▌                      | 465/1230 [3:30:52<5:31:05, 25.97s/it]
                                                                                
{'loss': 0.7631, 'grad_norm': 0.40429186820983887, 'learning_rate': 0.00013384254185758445, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.82, 'epoch': 0.33}

 38%|█████████████▌                      | 465/1230 [3:30:52<5:31:05, 25.97s/it]
 38%|█████████████▋                      | 466/1230 [3:31:22<5:43:25, 26.97s/it]
                                                                                
{'loss': 0.8019, 'grad_norm': 0.3995359241962433, 'learning_rate': 0.00013365002348452698, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.37, 'epoch': 0.33}

 38%|█████████████▋                      | 466/1230 [3:31:22<5:43:25, 26.97s/it]
 38%|█████████████▋                      | 467/1230 [3:31:48<5:39:44, 26.72s/it]
                                                                                
{'loss': 0.7547, 'grad_norm': 0.38715019822120667, 'learning_rate': 0.00013345726523128482, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.8, 'epoch': 0.33}

 38%|█████████████▋                      | 467/1230 [3:31:48<5:39:44, 26.72s/it]
 38%|█████████████▋                      | 468/1230 [3:32:13<5:32:10, 26.16s/it]
                                                                                
{'loss': 0.7566, 'grad_norm': 0.396226167678833, 'learning_rate': 0.00013326426843231266, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.49, 'epoch': 0.33}

 38%|█████████████▋                      | 468/1230 [3:32:13<5:32:10, 26.16s/it]
 38%|█████████████▋                      | 469/1230 [3:32:42<5:43:09, 27.06s/it]
                                                                                
{'loss': 0.765, 'grad_norm': 0.36862772703170776, 'learning_rate': 0.00013307103442371655, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.39, 'epoch': 0.33}

 38%|█████████████▋                      | 469/1230 [3:32:42<5:43:09, 27.06s/it]
 38%|█████████████▊                      | 470/1230 [3:33:07<5:34:51, 26.44s/it]
                                                                                
{'loss': 0.7654, 'grad_norm': 0.3937393128871918, 'learning_rate': 0.00013287756454324475, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.68, 'epoch': 0.33}

 38%|█████████████▊                      | 470/1230 [3:33:07<5:34:51, 26.44s/it]
 38%|█████████████▊                      | 471/1230 [3:33:32<5:29:29, 26.05s/it]
                                                                                
{'loss': 0.7819, 'grad_norm': 0.3958071768283844, 'learning_rate': 0.0001326838601302786, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.47, 'epoch': 0.33}

 38%|█████████████▊                      | 471/1230 [3:33:32<5:29:29, 26.05s/it]
 38%|█████████████▊                      | 472/1230 [3:34:02<5:42:28, 27.11s/it]
                                                                                
{'loss': 0.7622, 'grad_norm': 0.3987394869327545, 'learning_rate': 0.0001324899225258229, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.15, 'epoch': 0.33}

 38%|█████████████▊                      | 472/1230 [3:34:02<5:42:28, 27.11s/it]
 38%|█████████████▊                      | 473/1230 [3:34:31<5:51:49, 27.89s/it]
                                                                                
{'loss': 0.8043, 'grad_norm': 0.4056817591190338, 'learning_rate': 0.00013229575307249682, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.18, 'epoch': 0.33}

 38%|█████████████▊                      | 473/1230 [3:34:31<5:51:49, 27.89s/it]
 39%|█████████████▊                      | 474/1230 [3:35:01<5:56:42, 28.31s/it]
                                                                                
{'loss': 0.7639, 'grad_norm': 0.40141454339027405, 'learning_rate': 0.00013210135311452482, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.55, 'epoch': 0.34}

 39%|█████████████▊                      | 474/1230 [3:35:01<5:56:42, 28.31s/it]
 39%|█████████████▉                      | 475/1230 [3:35:25<5:43:36, 27.31s/it]
                                                                                
{'loss': 0.7358, 'grad_norm': 0.4025418162345886, 'learning_rate': 0.00013190672399772694, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.63, 'epoch': 0.34}

 39%|█████████████▉                      | 475/1230 [3:35:25<5:43:36, 27.31s/it]
 39%|█████████████▉                      | 476/1230 [3:35:55<5:49:50, 27.84s/it]
                                                                                
{'loss': 0.8107, 'grad_norm': 0.41207534074783325, 'learning_rate': 0.00013171186706950968, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.7, 'epoch': 0.34}

 39%|█████████████▉                      | 476/1230 [3:35:55<5:49:50, 27.84s/it]
 39%|█████████████▉                      | 477/1230 [3:36:19<5:38:17, 26.96s/it]
                                                                                
{'loss': 0.7513, 'grad_norm': 0.3980991840362549, 'learning_rate': 0.00013151678367885676, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.11, 'epoch': 0.34}

 39%|█████████████▉                      | 477/1230 [3:36:19<5:38:17, 26.96s/it]
 39%|█████████████▉                      | 478/1230 [3:36:45<5:31:22, 26.44s/it]
                                                                                
{'loss': 0.7936, 'grad_norm': 0.38969099521636963, 'learning_rate': 0.00013132147517631962, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.17, 'epoch': 0.34}

 39%|█████████████▉                      | 478/1230 [3:36:45<5:31:22, 26.44s/it]
 39%|██████████████                      | 479/1230 [3:37:10<5:26:22, 26.08s/it]
                                                                                
{'loss': 0.7617, 'grad_norm': 0.40378716588020325, 'learning_rate': 0.0001311259429140082, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.74, 'epoch': 0.34}

 39%|██████████████                      | 479/1230 [3:37:10<5:26:22, 26.08s/it]
 39%|██████████████                      | 480/1230 [3:37:39<5:37:56, 27.04s/it]
                                                                                
{'loss': 0.7401, 'grad_norm': 0.3645806908607483, 'learning_rate': 0.00013093018824558142, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.05, 'epoch': 0.34}

 39%|██████████████                      | 480/1230 [3:37:39<5:37:56, 27.04s/it]
 39%|██████████████                      | 481/1230 [3:38:05<5:32:16, 26.62s/it]
                                                                                
{'loss': 0.782, 'grad_norm': 0.3965825140476227, 'learning_rate': 0.00013073421252623795, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.81, 'epoch': 0.34}

 39%|██████████████                      | 481/1230 [3:38:05<5:32:16, 26.62s/it]
 39%|██████████████                      | 482/1230 [3:38:35<5:43:30, 27.55s/it]
                                                                                
{'loss': 0.7688, 'grad_norm': 0.373762309551239, 'learning_rate': 0.00013053801711270682, 'memory/max_active (GiB)': 8.23, 'memory/max_allocated (GiB)': 8.23, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.57, 'epoch': 0.34}

 39%|██████████████                      | 482/1230 [3:38:35<5:43:30, 27.55s/it]
 39%|██████████████▏                     | 483/1230 [3:39:04<5:51:08, 28.20s/it]
                                                                                
{'loss': 0.8092, 'grad_norm': 0.41407766938209534, 'learning_rate': 0.000130341603363238, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.12, 'epoch': 0.34}

 39%|██████████████▏                     | 483/1230 [3:39:04<5:51:08, 28.20s/it]
 39%|██████████████▏                     | 484/1230 [3:39:33<5:54:01, 28.47s/it]
                                                                                
{'loss': 0.7642, 'grad_norm': 0.38789892196655273, 'learning_rate': 0.00013014497263759287, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.86, 'epoch': 0.34}

 39%|██████████████▏                     | 484/1230 [3:39:33<5:54:01, 28.47s/it]
 39%|██████████████▏                     | 485/1230 [3:40:02<5:55:27, 28.63s/it]
                                                                                
{'loss': 0.8036, 'grad_norm': 0.4258516728878021, 'learning_rate': 0.00012994812629703507, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.32, 'epoch': 0.34}

 39%|██████████████▏                     | 485/1230 [3:40:02<5:55:27, 28.63s/it]
 40%|██████████████▏                     | 486/1230 [3:40:27<5:41:05, 27.51s/it]
                                                                                
{'loss': 0.7621, 'grad_norm': 0.4061269164085388, 'learning_rate': 0.00012975106570432083, 'memory/max_active (GiB)': 9.76, 'memory/max_allocated (GiB)': 9.76, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.01, 'epoch': 0.34}

 40%|██████████████▏                     | 486/1230 [3:40:27<5:41:05, 27.51s/it]
 40%|██████████████▎                     | 487/1230 [3:40:53<5:34:54, 27.04s/it]
                                                                                
{'loss': 0.7648, 'grad_norm': 0.3933376967906952, 'learning_rate': 0.00012955379222368967, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.03, 'epoch': 0.34}

 40%|██████████████▎                     | 487/1230 [3:40:53<5:34:54, 27.04s/it]
 40%|██████████████▎                     | 488/1230 [3:41:19<5:28:33, 26.57s/it]
                                                                                
{'loss': 0.7726, 'grad_norm': 0.3970009684562683, 'learning_rate': 0.00012935630722085492, 'memory/max_active (GiB)': 8.23, 'memory/max_allocated (GiB)': 8.23, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 212.48, 'epoch': 0.35}

 40%|██████████████▎                     | 488/1230 [3:41:19<5:28:33, 26.57s/it]
 40%|██████████████▎                     | 489/1230 [3:41:45<5:26:54, 26.47s/it]
                                                                                
{'loss': 0.7373, 'grad_norm': 0.41773009300231934, 'learning_rate': 0.00012915861206299427, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.86, 'epoch': 0.35}

 40%|██████████████▎                     | 489/1230 [3:41:45<5:26:54, 26.47s/it]
 40%|██████████████▎                     | 490/1230 [3:42:15<5:38:07, 27.42s/it]
                                                                                
{'loss': 0.775, 'grad_norm': 0.4052152931690216, 'learning_rate': 0.0001289607081187403, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.23, 'epoch': 0.35}

 40%|██████████████▎                     | 490/1230 [3:42:15<5:38:07, 27.42s/it]
 40%|██████████████▎                     | 491/1230 [3:42:40<5:29:56, 26.79s/it]
                                                                                
{'loss': 0.7314, 'grad_norm': 0.41575759649276733, 'learning_rate': 0.00012876259675817102, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.77, 'epoch': 0.35}

 40%|██████████████▎                     | 491/1230 [3:42:40<5:29:56, 26.79s/it]
 40%|██████████████▍                     | 492/1230 [3:43:10<5:40:13, 27.66s/it]
                                                                                
{'loss': 0.8013, 'grad_norm': 0.4295899569988251, 'learning_rate': 0.00012856427935280032, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.17, 'epoch': 0.35}

 40%|██████████████▍                     | 492/1230 [3:43:10<5:40:13, 27.66s/it]
 40%|██████████████▍                     | 493/1230 [3:43:35<5:31:15, 26.97s/it]
                                                                                
{'loss': 0.7004, 'grad_norm': 0.40650230646133423, 'learning_rate': 0.0001283657572755686, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.61, 'epoch': 0.35}

 40%|██████████████▍                     | 493/1230 [3:43:35<5:31:15, 26.97s/it]
 40%|██████████████▍                     | 494/1230 [3:44:00<5:25:01, 26.50s/it]
                                                                                
{'loss': 0.7763, 'grad_norm': 0.43630149960517883, 'learning_rate': 0.00012816703190083313, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.24, 'epoch': 0.35}

 40%|██████████████▍                     | 494/1230 [3:44:00<5:25:01, 26.50s/it]
 40%|██████████████▍                     | 495/1230 [3:44:26<5:20:34, 26.17s/it]
                                                                                
{'loss': 0.7692, 'grad_norm': 0.4215308129787445, 'learning_rate': 0.00012796810460435863, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.45, 'epoch': 0.35}

 40%|██████████████▍                     | 495/1230 [3:44:26<5:20:34, 26.17s/it]
 40%|██████████████▌                     | 496/1230 [3:44:51<5:16:45, 25.89s/it]
                                                                                
{'loss': 0.785, 'grad_norm': 0.42226913571357727, 'learning_rate': 0.00012776897676330772, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.74, 'epoch': 0.35}

 40%|██████████████▌                     | 496/1230 [3:44:51<5:16:45, 25.89s/it]
 40%|██████████████▌                     | 497/1230 [3:45:16<5:13:06, 25.63s/it]
                                                                                
{'loss': 0.7891, 'grad_norm': 0.4319060146808624, 'learning_rate': 0.00012756964975623133, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.22, 'epoch': 0.35}

 40%|██████████████▌                     | 497/1230 [3:45:16<5:13:06, 25.63s/it]
 40%|██████████████▌                     | 498/1230 [3:45:41<5:10:39, 25.46s/it]
                                                                                
{'loss': 0.748, 'grad_norm': 0.3773418366909027, 'learning_rate': 0.0001273701249630593, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.62, 'epoch': 0.35}

 40%|██████████████▌                     | 498/1230 [3:45:41<5:10:39, 25.46s/it]
 41%|██████████████▌                     | 499/1230 [3:46:11<5:25:11, 26.69s/it]
                                                                                
{'loss': 0.7669, 'grad_norm': 0.3841691315174103, 'learning_rate': 0.00012717040376509058, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.35, 'epoch': 0.35}

 41%|██████████████▌                     | 499/1230 [3:46:11<5:25:11, 26.69s/it]
 41%|██████████████▋                     | 500/1230 [3:46:36<5:19:32, 26.26s/it]
                                                                                
{'loss': 0.7974, 'grad_norm': 0.4163007140159607, 'learning_rate': 0.000126970487544984, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.16, 'epoch': 0.35}

 41%|██████████████▋                     | 500/1230 [3:46:36<5:19:32, 26.26s/it]
 41%|██████████████▋                     | 501/1230 [3:46:58<5:05:11, 25.12s/it]
                                                                                
{'loss': 0.8672, 'grad_norm': 1.7799453735351562, 'learning_rate': 0.0001267703776867484, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.07, 'epoch': 0.35}

 41%|██████████████▋                     | 501/1230 [3:46:58<5:05:11, 25.12s/it]
 41%|██████████████▋                     | 502/1230 [3:47:23<5:03:41, 25.03s/it]
                                                                                
{'loss': 0.7654, 'grad_norm': 0.4407631754875183, 'learning_rate': 0.00012657007557573315, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.98, 'epoch': 0.36}

 41%|██████████████▋                     | 502/1230 [3:47:23<5:03:41, 25.03s/it]
 41%|██████████████▋                     | 503/1230 [3:47:49<5:07:25, 25.37s/it]
                                                                                
{'loss': 0.7507, 'grad_norm': 0.3822515904903412, 'learning_rate': 0.00012636958259861872, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.83, 'epoch': 0.36}

 41%|██████████████▋                     | 503/1230 [3:47:49<5:07:25, 25.37s/it]
 41%|██████████████▊                     | 504/1230 [3:48:18<5:20:04, 26.45s/it]
                                                                                
{'loss': 0.7224, 'grad_norm': 0.3961871564388275, 'learning_rate': 0.0001261689001434067, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.11, 'epoch': 0.36}

 41%|██████████████▊                     | 504/1230 [3:48:18<5:20:04, 26.45s/it]
 41%|██████████████▊                     | 505/1230 [3:48:43<5:14:10, 26.00s/it]
                                                                                
{'loss': 0.7683, 'grad_norm': 0.40184205770492554, 'learning_rate': 0.00012596802959941066, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.6, 'epoch': 0.36}

 41%|██████████████▊                     | 505/1230 [3:48:43<5:14:10, 26.00s/it]
 41%|██████████████▊                     | 506/1230 [3:49:13<5:26:34, 27.06s/it]
                                                                                
{'loss': 0.8022, 'grad_norm': 0.4337049722671509, 'learning_rate': 0.00012576697235724618, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.73, 'epoch': 0.36}

 41%|██████████████▊                     | 506/1230 [3:49:13<5:26:34, 27.06s/it]
 41%|██████████████▊                     | 507/1230 [3:49:43<5:35:41, 27.86s/it]
                                                                                
{'loss': 0.7926, 'grad_norm': 0.4172649681568146, 'learning_rate': 0.00012556572980882135, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.84, 'epoch': 0.36}

 41%|██████████████▊                     | 507/1230 [3:49:43<5:35:41, 27.86s/it]
 41%|██████████████▊                     | 508/1230 [3:50:12<5:41:01, 28.34s/it]
                                                                                
{'loss': 0.7643, 'grad_norm': 0.4260568916797638, 'learning_rate': 0.00012536430334732714, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.41, 'epoch': 0.36}

 41%|██████████████▊                     | 508/1230 [3:50:12<5:41:01, 28.34s/it]
 41%|██████████████▉                     | 509/1230 [3:50:37<5:28:22, 27.33s/it]
                                                                                
{'loss': 0.7714, 'grad_norm': 0.41883915662765503, 'learning_rate': 0.00012516269436722776, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.72, 'epoch': 0.36}

 41%|██████████████▉                     | 509/1230 [3:50:37<5:28:22, 27.33s/it]
 41%|██████████████▉                     | 510/1230 [3:51:06<5:34:39, 27.89s/it]
                                                                                
{'loss': 0.7847, 'grad_norm': 0.4366205334663391, 'learning_rate': 0.00012496090426425093, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.17, 'epoch': 0.36}

 41%|██████████████▉                     | 510/1230 [3:51:06<5:34:39, 27.89s/it]
 42%|██████████████▉                     | 511/1230 [3:51:35<5:39:22, 28.32s/it]
                                                                                
{'loss': 0.7943, 'grad_norm': 0.40793347358703613, 'learning_rate': 0.0001247589344353783, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.15, 'epoch': 0.36}

 42%|██████████████▉                     | 511/1230 [3:51:35<5:39:22, 28.32s/it]
 42%|██████████████▉                     | 512/1230 [3:52:05<5:42:39, 28.63s/it]
                                                                                
{'loss': 0.7819, 'grad_norm': 0.4422757923603058, 'learning_rate': 0.00012455678627883577, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.06, 'epoch': 0.36}

 42%|██████████████▉                     | 512/1230 [3:52:05<5:42:39, 28.63s/it]
 42%|███████████████                     | 513/1230 [3:52:34<5:44:50, 28.86s/it]
                                                                                
{'loss': 0.7996, 'grad_norm': 0.39789584279060364, 'learning_rate': 0.00012435446119408381, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.13, 'epoch': 0.36}

 42%|███████████████                     | 513/1230 [3:52:34<5:44:50, 28.86s/it]
 42%|███████████████                     | 514/1230 [3:53:00<5:31:41, 27.80s/it]
                                                                                
{'loss': 0.7843, 'grad_norm': 0.41801339387893677, 'learning_rate': 0.0001241519605818077, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.8, 'epoch': 0.36}

 42%|███████████████                     | 514/1230 [3:53:00<5:31:41, 27.80s/it]
 42%|███████████████                     | 515/1230 [3:53:25<5:22:42, 27.08s/it]
                                                                                
{'loss': 0.7298, 'grad_norm': 0.4355681538581848, 'learning_rate': 0.00012394928584390789, 'memory/max_active (GiB)': 11.54, 'memory/max_allocated (GiB)': 11.54, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.76, 'epoch': 0.36}

 42%|███████████████                     | 515/1230 [3:53:25<5:22:42, 27.08s/it]
 42%|███████████████                     | 516/1230 [3:53:54<5:29:05, 27.66s/it]
                                                                                
{'loss': 0.7526, 'grad_norm': 0.3959862291812897, 'learning_rate': 0.00012374643838349033, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.87, 'epoch': 0.37}

 42%|███████████████                     | 516/1230 [3:53:54<5:29:05, 27.66s/it]
 42%|███████████████▏                    | 517/1230 [3:54:19<5:19:40, 26.90s/it]
                                                                                
{'loss': 0.7476, 'grad_norm': 0.42055782675743103, 'learning_rate': 0.00012354341960485669, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.44, 'epoch': 0.37}

 42%|███████████████▏                    | 517/1230 [3:54:19<5:19:40, 26.90s/it]
 42%|███████████████▏                    | 518/1230 [3:54:44<5:12:46, 26.36s/it]
                                                                                
{'loss': 0.7722, 'grad_norm': 0.4050011932849884, 'learning_rate': 0.00012334023091349475, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.89, 'epoch': 0.37}

 42%|███████████████▏                    | 518/1230 [3:54:44<5:12:46, 26.36s/it]
 42%|███████████████▏                    | 519/1230 [3:55:13<5:22:00, 27.17s/it]
                                                                                
{'loss': 0.7337, 'grad_norm': 0.39618802070617676, 'learning_rate': 0.0001231368737160684, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.95, 'epoch': 0.37}

 42%|███████████████▏                    | 519/1230 [3:55:13<5:22:00, 27.17s/it]
 42%|███████████████▏                    | 520/1230 [3:55:38<5:13:50, 26.52s/it]
                                                                                
{'loss': 0.7356, 'grad_norm': 0.44217830896377563, 'learning_rate': 0.0001229333494204083, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.87, 'epoch': 0.37}

 42%|███████████████▏                    | 520/1230 [3:55:38<5:13:50, 26.52s/it]
 42%|███████████████▏                    | 521/1230 [3:56:04<5:11:28, 26.36s/it]
                                                                                
{'loss': 0.7324, 'grad_norm': 0.4074406921863556, 'learning_rate': 0.00012272965943550175, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.21, 'epoch': 0.37}

 42%|███████████████▏                    | 521/1230 [3:56:04<5:11:28, 26.36s/it]
 42%|███████████████▎                    | 522/1230 [3:56:33<5:20:39, 27.17s/it]
                                                                                
{'loss': 0.74, 'grad_norm': 0.3830181658267975, 'learning_rate': 0.00012252580517148324, 'memory/max_active (GiB)': 8.86, 'memory/max_allocated (GiB)': 8.86, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.98, 'epoch': 0.37}

 42%|███████████████▎                    | 522/1230 [3:56:33<5:20:39, 27.17s/it]
 43%|███████████████▎                    | 523/1230 [3:57:03<5:29:23, 27.95s/it]
                                                                                
{'loss': 0.7759, 'grad_norm': 0.4161931276321411, 'learning_rate': 0.0001223217880396245, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.59, 'epoch': 0.37}

 43%|███████████████▎                    | 523/1230 [3:57:03<5:29:23, 27.95s/it]
 43%|███████████████▎                    | 524/1230 [3:57:33<5:34:14, 28.41s/it]
                                                                                
{'loss': 0.7595, 'grad_norm': 0.4144625961780548, 'learning_rate': 0.00012211760945232472, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.35, 'epoch': 0.37}

 43%|███████████████▎                    | 524/1230 [3:57:33<5:34:14, 28.41s/it]
 43%|███████████████▎                    | 525/1230 [3:58:02<5:37:29, 28.72s/it]
                                                                                
{'loss': 0.7734, 'grad_norm': 0.3830767571926117, 'learning_rate': 0.00012191327082310097, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.36, 'epoch': 0.37}

 43%|███████████████▎                    | 525/1230 [3:58:02<5:37:29, 28.72s/it]
 43%|███████████████▍                    | 526/1230 [3:58:31<5:39:38, 28.95s/it]
                                                                                
{'loss': 0.7793, 'grad_norm': 0.42880895733833313, 'learning_rate': 0.00012170877356657818, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.83, 'epoch': 0.37}

 43%|███████████████▍                    | 526/1230 [3:58:31<5:39:38, 28.95s/it]
 43%|███████████████▍                    | 527/1230 [3:59:01<5:39:45, 29.00s/it]
                                                                                
{'loss': 0.7444, 'grad_norm': 0.37180811166763306, 'learning_rate': 0.00012150411909847946, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.31, 'epoch': 0.37}

 43%|███████████████▍                    | 527/1230 [3:59:01<5:39:45, 29.00s/it]
 43%|███████████████▍                    | 528/1230 [3:59:26<5:25:22, 27.81s/it]
                                                                                
{'loss': 0.7559, 'grad_norm': 0.40364912152290344, 'learning_rate': 0.00012129930883561633, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.97, 'epoch': 0.37}

 43%|███████████████▍                    | 528/1230 [3:59:26<5:25:22, 27.81s/it]
 43%|███████████████▍                    | 529/1230 [3:59:55<5:31:07, 28.34s/it]
                                                                                
{'loss': 0.7513, 'grad_norm': 0.392648309469223, 'learning_rate': 0.00012109434419587887, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.04, 'epoch': 0.37}

 43%|███████████████▍                    | 529/1230 [3:59:55<5:31:07, 28.34s/it]
 43%|███████████████▌                    | 530/1230 [4:00:25<5:34:36, 28.68s/it]
                                                                                
{'loss': 0.7804, 'grad_norm': 0.430370956659317, 'learning_rate': 0.00012088922659822583, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.32, 'epoch': 0.38}

 43%|███████████████▌                    | 530/1230 [4:00:25<5:34:36, 28.68s/it]
 43%|███████████████▌                    | 531/1230 [4:00:50<5:22:13, 27.66s/it]
                                                                                
{'loss': 0.7234, 'grad_norm': 0.3841852843761444, 'learning_rate': 0.00012068395746267492, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.73, 'epoch': 0.38}

 43%|███████████████▌                    | 531/1230 [4:00:50<5:22:13, 27.66s/it]
 43%|███████████████▌                    | 532/1230 [4:01:16<5:16:09, 27.18s/it]
                                                                                
{'loss': 0.7034, 'grad_norm': 0.39239072799682617, 'learning_rate': 0.00012047853821029299, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.73, 'epoch': 0.38}

 43%|███████████████▌                    | 532/1230 [4:01:16<5:16:09, 27.18s/it]
 43%|███████████████▌                    | 533/1230 [4:01:41<5:08:49, 26.58s/it]
                                                                                
{'loss': 0.7498, 'grad_norm': 0.40071728825569153, 'learning_rate': 0.00012027297026318606, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.93, 'epoch': 0.38}

 43%|███████████████▌                    | 533/1230 [4:01:41<5:08:49, 26.58s/it]
 43%|███████████████▋                    | 534/1230 [4:02:06<5:03:46, 26.19s/it]
                                                                                
{'loss': 0.7444, 'grad_norm': 0.40049320459365845, 'learning_rate': 0.00012006725504448961, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.9, 'epoch': 0.38}

 43%|███████████████▋                    | 534/1230 [4:02:06<5:03:46, 26.19s/it]
 43%|███████████████▋                    | 535/1230 [4:02:36<5:14:38, 27.16s/it]
                                                                                
{'loss': 0.7437, 'grad_norm': 0.3714231848716736, 'learning_rate': 0.00011986139397835867, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.99, 'epoch': 0.38}

 43%|███████████████▋                    | 535/1230 [4:02:36<5:14:38, 27.16s/it]
 44%|███████████████▋                    | 536/1230 [4:03:02<5:09:21, 26.75s/it]
                                                                                
{'loss': 0.7377, 'grad_norm': 0.4161554276943207, 'learning_rate': 0.00011965538848995786, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.72, 'epoch': 0.38}

 44%|███████████████▋                    | 536/1230 [4:03:02<5:09:21, 26.75s/it]
 44%|███████████████▋                    | 537/1230 [4:03:31<5:18:06, 27.54s/it]
                                                                                
{'loss': 0.7384, 'grad_norm': 0.42263901233673096, 'learning_rate': 0.00011944924000545182, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.24, 'epoch': 0.38}

 44%|███████████████▋                    | 537/1230 [4:03:31<5:18:06, 27.54s/it]
 44%|███████████████▋                    | 538/1230 [4:03:56<5:09:45, 26.86s/it]
                                                                                
{'loss': 0.7293, 'grad_norm': 0.39381951093673706, 'learning_rate': 0.000119242949951995, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.57, 'epoch': 0.38}

 44%|███████████████▋                    | 538/1230 [4:03:56<5:09:45, 26.86s/it]
 44%|███████████████▊                    | 539/1230 [4:04:21<5:02:41, 26.28s/it]
                                                                                
{'loss': 0.7171, 'grad_norm': 0.38888514041900635, 'learning_rate': 0.00011903651975772195, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.01, 'epoch': 0.38}

 44%|███████████████▊                    | 539/1230 [4:04:21<5:02:41, 26.28s/it]
 44%|███████████████▊                    | 540/1230 [4:04:46<4:57:49, 25.90s/it]
                                                                                
{'loss': 0.7551, 'grad_norm': 0.3880782723426819, 'learning_rate': 0.0001188299508517374, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.71, 'epoch': 0.38}

 44%|███████████████▊                    | 540/1230 [4:04:46<4:57:49, 25.90s/it]
 44%|███████████████▊                    | 541/1230 [4:05:15<5:08:32, 26.87s/it]
                                                                                
{'loss': 0.7664, 'grad_norm': 0.4123733937740326, 'learning_rate': 0.00011862324466410644, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.81, 'epoch': 0.38}

 44%|███████████████▊                    | 541/1230 [4:05:15<5:08:32, 26.87s/it]
 44%|███████████████▊                    | 542/1230 [4:05:41<5:04:03, 26.52s/it]
                                                                                
{'loss': 0.757, 'grad_norm': 0.413076788187027, 'learning_rate': 0.00011841640262584452, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.83, 'epoch': 0.38}

 44%|███████████████▊                    | 542/1230 [4:05:41<5:04:03, 26.52s/it]
 44%|███████████████▉                    | 543/1230 [4:06:10<5:13:26, 27.38s/it]
                                                                                
{'loss': 0.7801, 'grad_norm': 0.38010311126708984, 'learning_rate': 0.0001182094261689075, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.55, 'epoch': 0.38}

 44%|███████████████▉                    | 543/1230 [4:06:10<5:13:26, 27.38s/it]
 44%|███████████████▉                    | 544/1230 [4:06:40<5:19:01, 27.90s/it]
                                                                                
{'loss': 0.7506, 'grad_norm': 0.38604605197906494, 'learning_rate': 0.00011800231672618189, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.81, 'epoch': 0.38}

 44%|███████████████▉                    | 544/1230 [4:06:40<5:19:01, 27.90s/it]
 44%|███████████████▉                    | 545/1230 [4:07:05<5:09:25, 27.10s/it]
                                                                                
{'loss': 0.768, 'grad_norm': 0.41267746686935425, 'learning_rate': 0.00011779507573147488, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.25, 'epoch': 0.39}

 44%|███████████████▉                    | 545/1230 [4:07:05<5:09:25, 27.10s/it]
 44%|███████████████▉                    | 546/1230 [4:07:31<5:05:43, 26.82s/it]
                                                                                
{'loss': 0.7548, 'grad_norm': 0.39052876830101013, 'learning_rate': 0.00011758770461950429, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.59, 'epoch': 0.39}

 44%|███████████████▉                    | 546/1230 [4:07:31<5:05:43, 26.82s/it]
 44%|████████████████                    | 547/1230 [4:08:00<5:14:15, 27.61s/it]
                                                                                
{'loss': 0.7857, 'grad_norm': 0.3931976854801178, 'learning_rate': 0.00011738020482588881, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.86, 'epoch': 0.39}

 44%|████████████████                    | 547/1230 [4:08:00<5:14:15, 27.61s/it]
 45%|████████████████                    | 548/1230 [4:08:26<5:06:53, 27.00s/it]
                                                                                
{'loss': 0.7625, 'grad_norm': 0.4198307693004608, 'learning_rate': 0.00011717257778713801, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.03, 'epoch': 0.39}

 45%|████████████████                    | 548/1230 [4:08:26<5:06:53, 27.00s/it]
 45%|████████████████                    | 549/1230 [4:08:52<5:03:39, 26.75s/it]
                                                                                
{'loss': 0.7732, 'grad_norm': 0.37891459465026855, 'learning_rate': 0.00011696482494064225, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.65, 'epoch': 0.39}

 45%|████████████████                    | 549/1230 [4:08:52<5:03:39, 26.75s/it]
 45%|████████████████                    | 550/1230 [4:09:21<5:11:29, 27.48s/it]
                                                                                
{'loss': 0.7572, 'grad_norm': 0.3719899356365204, 'learning_rate': 0.00011675694772466302, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.09, 'epoch': 0.39}

 45%|████████████████                    | 550/1230 [4:09:21<5:11:29, 27.48s/it]
 45%|████████████████▏                   | 551/1230 [4:09:50<5:16:09, 27.94s/it]
                                                                                
{'loss': 0.7846, 'grad_norm': 0.3879837393760681, 'learning_rate': 0.00011654894757832263, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.64, 'epoch': 0.39}

 45%|████████████████▏                   | 551/1230 [4:09:50<5:16:09, 27.94s/it]
 45%|████████████████▏                   | 552/1230 [4:10:20<5:20:06, 28.33s/it]
                                                                                
{'loss': 0.7338, 'grad_norm': 0.3644126057624817, 'learning_rate': 0.00011634082594159456, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.48, 'epoch': 0.39}

 45%|████████████████▏                   | 552/1230 [4:10:20<5:20:06, 28.33s/it]
 45%|████████████████▏                   | 553/1230 [4:10:49<5:23:19, 28.66s/it]
                                                                                
{'loss': 0.7659, 'grad_norm': 0.38544681668281555, 'learning_rate': 0.00011613258425529335, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.89, 'epoch': 0.39}

 45%|████████████████▏                   | 553/1230 [4:10:49<5:23:19, 28.66s/it]
 45%|████████████████▏                   | 554/1230 [4:11:15<5:14:06, 27.88s/it]
                                                                                
{'loss': 0.764, 'grad_norm': 0.4077942371368408, 'learning_rate': 0.00011592422396106455, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.63, 'epoch': 0.39}

 45%|████████████████▏                   | 554/1230 [4:11:15<5:14:06, 27.88s/it]
 45%|████████████████▏                   | 555/1230 [4:11:45<5:19:24, 28.39s/it]
                                                                                
{'loss': 0.7733, 'grad_norm': 0.40184280276298523, 'learning_rate': 0.00011571574650137494, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.64, 'epoch': 0.39}

 45%|████████████████▏                   | 555/1230 [4:11:45<5:19:24, 28.39s/it]
 45%|████████████████▎                   | 556/1230 [4:12:14<5:22:13, 28.68s/it]
                                                                                
{'loss': 0.7719, 'grad_norm': 0.3854644298553467, 'learning_rate': 0.00011550715331950232, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.32, 'epoch': 0.39}

 45%|████████████████▎                   | 556/1230 [4:12:14<5:22:13, 28.68s/it]
 45%|████████████████▎                   | 557/1230 [4:12:39<5:09:45, 27.62s/it]
                                                                                
{'loss': 0.7283, 'grad_norm': 0.382046639919281, 'learning_rate': 0.00011529844585952574, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.38, 'epoch': 0.39}

 45%|████████████████▎                   | 557/1230 [4:12:39<5:09:45, 27.62s/it]
 45%|████████████████▎                   | 558/1230 [4:13:04<5:01:20, 26.90s/it]
                                                                                
{'loss': 0.752, 'grad_norm': 0.4347343146800995, 'learning_rate': 0.00011508962556631527, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.23, 'epoch': 0.39}

 45%|████████████████▎                   | 558/1230 [4:13:04<5:01:20, 26.90s/it]
 45%|████████████████▎                   | 559/1230 [4:13:34<5:09:39, 27.69s/it]
                                                                                
{'loss': 0.7754, 'grad_norm': 0.3942156732082367, 'learning_rate': 0.00011488069388552225, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.11, 'epoch': 0.4}

 45%|████████████████▎                   | 559/1230 [4:13:34<5:09:39, 27.69s/it]
 46%|████████████████▍                   | 560/1230 [4:13:59<5:01:13, 26.98s/it]
                                                                                
{'loss': 0.7951, 'grad_norm': 0.45355433225631714, 'learning_rate': 0.000114671652263569, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.06, 'epoch': 0.4}

 46%|████████████████▍                   | 560/1230 [4:13:59<5:01:13, 26.98s/it]
 46%|████████████████▍                   | 561/1230 [4:14:29<5:09:38, 27.77s/it]
                                                                                
{'loss': 0.7403, 'grad_norm': 0.406571626663208, 'learning_rate': 0.0001144625021476391, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.06, 'epoch': 0.4}

 46%|████████████████▍                   | 561/1230 [4:14:29<5:09:38, 27.77s/it]
 46%|████████████████▍                   | 562/1230 [4:14:54<5:00:20, 26.98s/it]
                                                                                
{'loss': 0.7382, 'grad_norm': 0.4273393154144287, 'learning_rate': 0.00011425324498566715, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.96, 'epoch': 0.4}

 46%|████████████████▍                   | 562/1230 [4:14:54<5:00:20, 26.98s/it]
 46%|████████████████▍                   | 563/1230 [4:15:20<4:55:51, 26.61s/it]
                                                                                
{'loss': 0.7784, 'grad_norm': 0.4163869619369507, 'learning_rate': 0.00011404388222632885, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.73, 'epoch': 0.4}

 46%|████████████████▍                   | 563/1230 [4:15:20<4:55:51, 26.61s/it]
 46%|████████████████▌                   | 564/1230 [4:15:45<4:50:20, 26.16s/it]
                                                                                
{'loss': 0.7, 'grad_norm': 0.392711877822876, 'learning_rate': 0.0001138344153190309, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.74, 'epoch': 0.4}

 46%|████████████████▌                   | 564/1230 [4:15:45<4:50:20, 26.16s/it]
 46%|████████████████▌                   | 565/1230 [4:16:10<4:47:04, 25.90s/it]
                                                                                
{'loss': 0.7153, 'grad_norm': 0.4328343868255615, 'learning_rate': 0.00011362484571390107, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.36, 'epoch': 0.4}

 46%|████████████████▌                   | 565/1230 [4:16:10<4:47:04, 25.90s/it]
 46%|████████████████▌                   | 566/1230 [4:16:40<4:58:17, 26.95s/it]
                                                                                
{'loss': 0.7399, 'grad_norm': 0.37563255429267883, 'learning_rate': 0.0001134151748617781, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.36, 'epoch': 0.4}

 46%|████████████████▌                   | 566/1230 [4:16:40<4:58:17, 26.95s/it]
 46%|████████████████▌                   | 567/1230 [4:17:05<4:52:24, 26.46s/it]
                                                                                
{'loss': 0.729, 'grad_norm': 0.3845216929912567, 'learning_rate': 0.00011320540421420158, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.99, 'epoch': 0.4}

 46%|████████████████▌                   | 567/1230 [4:17:05<4:52:24, 26.46s/it]
 46%|████████████████▌                   | 568/1230 [4:17:34<5:02:08, 27.38s/it]
                                                                                
{'loss': 0.7611, 'grad_norm': 0.42829522490501404, 'learning_rate': 0.00011299553522340207, 'memory/max_active (GiB)': 10.53, 'memory/max_allocated (GiB)': 10.53, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.15, 'epoch': 0.4}

 46%|████████████████▌                   | 568/1230 [4:17:34<5:02:08, 27.38s/it]
 46%|████████████████▋                   | 569/1230 [4:17:59<4:53:46, 26.67s/it]
                                                                                
{'loss': 0.7394, 'grad_norm': 0.42085736989974976, 'learning_rate': 0.0001127855693422909, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.57, 'epoch': 0.4}

 46%|████████████████▋                   | 569/1230 [4:17:59<4:53:46, 26.67s/it]
 46%|████████████████▋                   | 570/1230 [4:18:25<4:50:54, 26.45s/it]
                                                                                
{'loss': 0.8189, 'grad_norm': 0.46468648314476013, 'learning_rate': 0.00011257550802445018, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.83, 'epoch': 0.4}

 46%|████████████████▋                   | 570/1230 [4:18:25<4:50:54, 26.45s/it]
 46%|████████████████▋                   | 571/1230 [4:18:52<4:49:34, 26.37s/it]
                                                                                
{'loss': 0.7617, 'grad_norm': 0.40098556876182556, 'learning_rate': 0.00011236535272412275, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.69, 'epoch': 0.4}

 46%|████████████████▋                   | 571/1230 [4:18:52<4:49:34, 26.37s/it]
 47%|████████████████▋                   | 572/1230 [4:19:18<4:49:20, 26.38s/it]
                                                                                
{'loss': 0.7248, 'grad_norm': 0.3796180188655853, 'learning_rate': 0.00011215510489620205, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.13, 'epoch': 0.4}

 47%|████████████████▋                   | 572/1230 [4:19:18<4:49:20, 26.38s/it]
 47%|████████████████▊                   | 573/1230 [4:19:47<4:58:16, 27.24s/it]
                                                                                
{'loss': 0.7423, 'grad_norm': 0.3987162709236145, 'learning_rate': 0.00011194476599622207, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.2, 'epoch': 0.41}

 47%|████████████████▊                   | 573/1230 [4:19:47<4:58:16, 27.24s/it]
 47%|████████████████▊                   | 574/1230 [4:20:12<4:50:06, 26.53s/it]
                                                                                
{'loss': 0.7531, 'grad_norm': 0.4186418056488037, 'learning_rate': 0.00011173433748034734, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.88, 'epoch': 0.41}

 47%|████████████████▊                   | 574/1230 [4:20:12<4:50:06, 26.53s/it]
 47%|████████████████▊                   | 575/1230 [4:20:38<4:46:05, 26.21s/it]
                                                                                
{'loss': 0.7231, 'grad_norm': 0.38458582758903503, 'learning_rate': 0.00011152382080536273, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.31, 'epoch': 0.41}

 47%|████████████████▊                   | 575/1230 [4:20:38<4:46:05, 26.21s/it]
 47%|████████████████▊                   | 576/1230 [4:21:04<4:45:52, 26.23s/it]
                                                                                
{'loss': 0.7554, 'grad_norm': 0.3883051872253418, 'learning_rate': 0.00011131321742866352, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.31, 'epoch': 0.41}

 47%|████████████████▊                   | 576/1230 [4:21:04<4:45:52, 26.23s/it]
 47%|████████████████▉                   | 577/1230 [4:21:29<4:43:01, 26.01s/it]
                                                                                
{'loss': 0.7637, 'grad_norm': 0.41564318537712097, 'learning_rate': 0.00011110252880824515, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.28, 'epoch': 0.41}

 47%|████████████████▉                   | 577/1230 [4:21:29<4:43:01, 26.01s/it]
 47%|████████████████▉                   | 578/1230 [4:21:59<4:54:42, 27.12s/it]
                                                                                
{'loss': 0.7048, 'grad_norm': 0.38103336095809937, 'learning_rate': 0.0001108917564026932, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.78, 'epoch': 0.41}

 47%|████████████████▉                   | 578/1230 [4:21:59<4:54:42, 27.12s/it]
 47%|████████████████▉                   | 579/1230 [4:22:25<4:50:53, 26.81s/it]
                                                                                
{'loss': 0.7478, 'grad_norm': 0.42740583419799805, 'learning_rate': 0.00011068090167117332, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.89, 'epoch': 0.41}

 47%|████████████████▉                   | 579/1230 [4:22:25<4:50:53, 26.81s/it]
 47%|████████████████▉                   | 580/1230 [4:22:54<4:58:37, 27.57s/it]
                                                                                
{'loss': 0.8022, 'grad_norm': 0.4371357560157776, 'learning_rate': 0.00011046996607342109, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.21, 'epoch': 0.41}

 47%|████████████████▉                   | 580/1230 [4:22:54<4:58:37, 27.57s/it]
 47%|█████████████████                   | 581/1230 [4:23:24<5:04:04, 28.11s/it]
                                                                                
{'loss': 0.7335, 'grad_norm': 0.40495625138282776, 'learning_rate': 0.00011025895106973192, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.24, 'epoch': 0.41}

 47%|█████████████████                   | 581/1230 [4:23:24<5:04:04, 28.11s/it]
 47%|█████████████████                   | 582/1230 [4:23:49<4:53:51, 27.21s/it]
                                                                                
{'loss': 0.6996, 'grad_norm': 0.37762078642845154, 'learning_rate': 0.0001100478581209509, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.6, 'epoch': 0.41}

 47%|█████████████████                   | 582/1230 [4:23:49<4:53:51, 27.21s/it]
 47%|█████████████████                   | 583/1230 [4:24:14<4:46:33, 26.57s/it]
                                                                                
{'loss': 0.7424, 'grad_norm': 0.40330711007118225, 'learning_rate': 0.00010983668868846282, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.98, 'epoch': 0.41}

 47%|█████████████████                   | 583/1230 [4:24:14<4:46:33, 26.57s/it]
 47%|█████████████████                   | 584/1230 [4:24:43<4:54:54, 27.39s/it]
                                                                                
{'loss': 0.7852, 'grad_norm': 0.4002113342285156, 'learning_rate': 0.00010962544423418188, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.06, 'epoch': 0.41}

 47%|█████████████████                   | 584/1230 [4:24:43<4:54:54, 27.39s/it]
 48%|█████████████████                   | 585/1230 [4:25:08<4:47:09, 26.71s/it]
                                                                                
{'loss': 0.7434, 'grad_norm': 0.39790281653404236, 'learning_rate': 0.00010941412622054172, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.99, 'epoch': 0.41}

 48%|█████████████████                   | 585/1230 [4:25:08<4:47:09, 26.71s/it]
 48%|█████████████████▏                  | 586/1230 [4:25:38<4:54:48, 27.47s/it]
                                                                                
{'loss': 0.7552, 'grad_norm': 0.370268851518631, 'learning_rate': 0.00010920273611048514, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.65, 'epoch': 0.41}

 48%|█████████████████▏                  | 586/1230 [4:25:38<4:54:48, 27.47s/it]
 48%|█████████████████▏                  | 587/1230 [4:26:03<4:46:51, 26.77s/it]
                                                                                
{'loss': 0.758, 'grad_norm': 0.45418310165405273, 'learning_rate': 0.00010899127536745414, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.95, 'epoch': 0.42}

 48%|█████████████████▏                  | 587/1230 [4:26:03<4:46:51, 26.77s/it]
 48%|█████████████████▏                  | 588/1230 [4:26:32<4:54:42, 27.54s/it]
                                                                                
{'loss': 0.7299, 'grad_norm': 0.37882599234580994, 'learning_rate': 0.0001087797454553797, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.71, 'epoch': 0.42}

 48%|█████████████████▏                  | 588/1230 [4:26:32<4:54:42, 27.54s/it]
 48%|█████████████████▏                  | 589/1230 [4:27:02<5:00:17, 28.11s/it]
                                                                                
{'loss': 0.7528, 'grad_norm': 0.4103173315525055, 'learning_rate': 0.00010856814783867156, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.65, 'epoch': 0.42}

 48%|█████████████████▏                  | 589/1230 [4:27:02<5:00:17, 28.11s/it]
 48%|█████████████████▎                  | 590/1230 [4:27:27<4:51:31, 27.33s/it]
                                                                                
{'loss': 0.7756, 'grad_norm': 0.41566282510757446, 'learning_rate': 0.0001083564839822083, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.21, 'epoch': 0.42}

 48%|█████████████████▎                  | 590/1230 [4:27:27<4:51:31, 27.33s/it]
 48%|█████████████████▎                  | 591/1230 [4:27:53<4:46:58, 26.95s/it]
                                                                                
{'loss': 0.7153, 'grad_norm': 0.41338640451431274, 'learning_rate': 0.00010814475535132706, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.01, 'epoch': 0.42}

 48%|█████████████████▎                  | 591/1230 [4:27:53<4:46:58, 26.95s/it]
 48%|█████████████████▎                  | 592/1230 [4:28:23<4:54:10, 27.67s/it]
                                                                                
{'loss': 0.7516, 'grad_norm': 0.444596529006958, 'learning_rate': 0.00010793296341181328, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.05, 'epoch': 0.42}

 48%|█████████████████▎                  | 592/1230 [4:28:23<4:54:10, 27.67s/it]
 48%|█████████████████▎                  | 593/1230 [4:28:48<4:45:33, 26.90s/it]
                                                                                
{'loss': 0.7554, 'grad_norm': 0.41693365573883057, 'learning_rate': 0.00010772110962989085, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.68, 'epoch': 0.42}

 48%|█████████████████▎                  | 593/1230 [4:28:48<4:45:33, 26.90s/it]
 48%|█████████████████▍                  | 594/1230 [4:29:17<4:52:02, 27.55s/it]
                                                                                
{'loss': 0.7759, 'grad_norm': 0.38711750507354736, 'learning_rate': 0.00010750919547221168, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.37, 'epoch': 0.42}

 48%|█████████████████▍                  | 594/1230 [4:29:17<4:52:02, 27.55s/it]
 48%|█████████████████▍                  | 595/1230 [4:29:46<4:56:58, 28.06s/it]
                                                                                
{'loss': 0.7401, 'grad_norm': 0.37913456559181213, 'learning_rate': 0.00010729722240584568, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.94, 'epoch': 0.42}

 48%|█████████████████▍                  | 595/1230 [4:29:46<4:56:58, 28.06s/it]
 48%|█████████████████▍                  | 596/1230 [4:30:12<4:50:29, 27.49s/it]
                                                                                
{'loss': 0.7031, 'grad_norm': 0.4284683167934418, 'learning_rate': 0.00010708519189827064, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.07, 'epoch': 0.42}

 48%|█████████████████▍                  | 596/1230 [4:30:12<4:50:29, 27.49s/it]
 49%|█████████████████▍                  | 597/1230 [4:30:38<4:43:26, 26.87s/it]
                                                                                
{'loss': 0.7495, 'grad_norm': 0.4017777740955353, 'learning_rate': 0.00010687310541736191, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.08, 'epoch': 0.42}

 49%|█████████████████▍                  | 597/1230 [4:30:38<4:43:26, 26.87s/it]
 49%|█████████████████▌                  | 598/1230 [4:31:07<4:51:44, 27.70s/it]
                                                                                
{'loss': 0.7259, 'grad_norm': 0.38410329818725586, 'learning_rate': 0.00010666096443138245, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.2, 'epoch': 0.42}

 49%|█████████████████▌                  | 598/1230 [4:31:07<4:51:44, 27.70s/it]
 49%|█████████████████▌                  | 599/1230 [4:31:37<4:57:12, 28.26s/it]
                                                                                
{'loss': 0.7825, 'grad_norm': 0.39601102471351624, 'learning_rate': 0.00010644877040897248, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.27, 'epoch': 0.42}

 49%|█████████████████▌                  | 599/1230 [4:31:37<4:57:12, 28.26s/it]
 49%|█████████████████▌                  | 600/1230 [4:32:03<4:49:10, 27.54s/it]
                                                                                
{'loss': 0.7698, 'grad_norm': 0.416627436876297, 'learning_rate': 0.0001062365248191394, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.88, 'epoch': 0.42}

 49%|█████████████████▌                  | 600/1230 [4:32:03<4:49:10, 27.54s/it]
 49%|█████████████████▌                  | 601/1230 [4:32:31<4:52:59, 27.95s/it]
                                                                                
{'loss': 0.7697, 'grad_norm': 0.38799187541007996, 'learning_rate': 0.00010602422913124764, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.07, 'epoch': 0.43}

 49%|█████████████████▌                  | 601/1230 [4:32:31<4:52:59, 27.95s/it]
 49%|█████████████████▌                  | 602/1230 [4:32:58<4:46:57, 27.42s/it]
                                                                                
{'loss': 0.7814, 'grad_norm': 0.4259137213230133, 'learning_rate': 0.00010581188481500839, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.31, 'epoch': 0.43}

 49%|█████████████████▌                  | 602/1230 [4:32:58<4:46:57, 27.42s/it]
 49%|█████████████████▋                  | 603/1230 [4:33:27<4:52:52, 28.03s/it]
                                                                                
{'loss': 0.7365, 'grad_norm': 0.3838924169540405, 'learning_rate': 0.00010559949334046962, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.15, 'epoch': 0.43}

 49%|█████████████████▋                  | 603/1230 [4:33:27<4:52:52, 28.03s/it]
 49%|█████████████████▋                  | 604/1230 [4:33:56<4:56:29, 28.42s/it]
                                                                                
{'loss': 0.7487, 'grad_norm': 0.3951365649700165, 'learning_rate': 0.00010538705617800559, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.46, 'epoch': 0.43}

 49%|█████████████████▋                  | 604/1230 [4:33:56<4:56:29, 28.42s/it]
 49%|█████████████████▋                  | 605/1230 [4:34:22<4:46:21, 27.49s/it]
                                                                                
{'loss': 0.7693, 'grad_norm': 0.41876882314682007, 'learning_rate': 0.00010517457479830705, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.0, 'epoch': 0.43}

 49%|█████████████████▋                  | 605/1230 [4:34:22<4:46:21, 27.49s/it]
 49%|█████████████████▋                  | 606/1230 [4:34:52<4:53:16, 28.20s/it]
                                                                                
{'loss': 0.7623, 'grad_norm': 0.3912851810455322, 'learning_rate': 0.00010496205067237072, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.06, 'epoch': 0.43}

 49%|█████████████████▋                  | 606/1230 [4:34:52<4:53:16, 28.20s/it]
 49%|█████████████████▊                  | 607/1230 [4:35:17<4:43:56, 27.35s/it]
                                                                                
{'loss': 0.7469, 'grad_norm': 0.4032377004623413, 'learning_rate': 0.0001047494852714893, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.15, 'epoch': 0.43}

 49%|█████████████████▊                  | 607/1230 [4:35:17<4:43:56, 27.35s/it]
 49%|█████████████████▊                  | 608/1230 [4:35:42<4:37:04, 26.73s/it]
                                                                                
{'loss': 0.7936, 'grad_norm': 0.4218188226222992, 'learning_rate': 0.00010453688006724125, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.74, 'epoch': 0.43}

 49%|█████████████████▊                  | 608/1230 [4:35:42<4:37:04, 26.73s/it]
 50%|█████████████████▊                  | 609/1230 [4:36:12<4:44:36, 27.50s/it]
                                                                                
{'loss': 0.7645, 'grad_norm': 0.4416923224925995, 'learning_rate': 0.00010432423653148054, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.49, 'epoch': 0.43}

 50%|█████████████████▊                  | 609/1230 [4:36:12<4:44:36, 27.50s/it]
 50%|█████████████████▊                  | 610/1230 [4:36:41<4:50:24, 28.10s/it]
                                                                                
{'loss': 0.7661, 'grad_norm': 0.3789323568344116, 'learning_rate': 0.00010411155613632656, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.98, 'epoch': 0.43}

 50%|█████████████████▊                  | 610/1230 [4:36:41<4:50:24, 28.10s/it]
 50%|█████████████████▉                  | 611/1230 [4:37:11<4:54:16, 28.52s/it]
                                                                                
{'loss': 0.7573, 'grad_norm': 0.38653433322906494, 'learning_rate': 0.00010389884035415387, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.4, 'epoch': 0.43}

 50%|█████████████████▉                  | 611/1230 [4:37:11<4:54:16, 28.52s/it]
 50%|█████████████████▉                  | 612/1230 [4:37:36<4:44:33, 27.63s/it]
                                                                                
{'loss': 0.7552, 'grad_norm': 0.4137389361858368, 'learning_rate': 0.00010368609065758198, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.0, 'epoch': 0.43}

 50%|█████████████████▉                  | 612/1230 [4:37:36<4:44:33, 27.63s/it]
 50%|█████████████████▉                  | 613/1230 [4:38:01<4:37:02, 26.94s/it]
                                                                                
{'loss': 0.75, 'grad_norm': 0.4579874873161316, 'learning_rate': 0.00010347330851946516, 'memory/max_active (GiB)': 11.04, 'memory/max_allocated (GiB)': 11.04, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.52, 'epoch': 0.43}

 50%|█████████████████▉                  | 613/1230 [4:38:01<4:37:02, 26.94s/it]
 50%|█████████████████▉                  | 614/1230 [4:38:27<4:32:33, 26.55s/it]
                                                                                
{'loss': 0.7711, 'grad_norm': 0.4558424651622772, 'learning_rate': 0.0001032604954128824, 'memory/max_active (GiB)': 8.86, 'memory/max_allocated (GiB)': 8.86, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.11, 'epoch': 0.43}

 50%|█████████████████▉                  | 614/1230 [4:38:27<4:32:33, 26.55s/it]
 50%|██████████████████                  | 615/1230 [4:38:57<4:41:05, 27.42s/it]
                                                                                
{'loss': 0.7554, 'grad_norm': 0.3914186954498291, 'learning_rate': 0.00010304765281112691, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.85, 'epoch': 0.44}

 50%|██████████████████                  | 615/1230 [4:38:57<4:41:05, 27.42s/it]
 50%|██████████████████                  | 616/1230 [4:39:25<4:45:11, 27.87s/it]
                                                                                
{'loss': 0.713, 'grad_norm': 0.36984920501708984, 'learning_rate': 0.0001028347821876962, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.01, 'epoch': 0.44}

 50%|██████████████████                  | 616/1230 [4:39:25<4:45:11, 27.87s/it]
 50%|██████████████████                  | 617/1230 [4:39:54<4:48:15, 28.22s/it]
                                                                                
{'loss': 0.7388, 'grad_norm': 0.3957446813583374, 'learning_rate': 0.0001026218850162818, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.02, 'epoch': 0.44}

 50%|██████████████████                  | 617/1230 [4:39:54<4:48:15, 28.22s/it]
 50%|██████████████████                  | 618/1230 [4:40:24<4:51:17, 28.56s/it]
                                                                                
{'loss': 0.7784, 'grad_norm': 0.4001215994358063, 'learning_rate': 0.00010240896277075895, 'memory/max_active (GiB)': 9.76, 'memory/max_allocated (GiB)': 9.76, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.7, 'epoch': 0.44}

 50%|██████████████████                  | 618/1230 [4:40:24<4:51:17, 28.56s/it]
 50%|██████████████████                  | 619/1230 [4:40:53<4:52:32, 28.73s/it]
                                                                                
{'loss': 0.7509, 'grad_norm': 0.37769433856010437, 'learning_rate': 0.00010219601692517649, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.65, 'epoch': 0.44}

 50%|██████████████████                  | 619/1230 [4:40:53<4:52:32, 28.73s/it]
 50%|██████████████████▏                 | 620/1230 [4:41:22<4:54:08, 28.93s/it]
                                                                                
{'loss': 0.7422, 'grad_norm': 0.38542476296424866, 'learning_rate': 0.00010198304895374667, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.35, 'epoch': 0.44}

 50%|██████████████████▏                 | 620/1230 [4:41:22<4:54:08, 28.93s/it]
 50%|██████████████████▏                 | 621/1230 [4:41:52<4:55:08, 29.08s/it]
                                                                                
{'loss': 0.7432, 'grad_norm': 0.39301788806915283, 'learning_rate': 0.00010177006033083494, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.17, 'epoch': 0.44}

 50%|██████████████████▏                 | 621/1230 [4:41:52<4:55:08, 29.08s/it]
 51%|██████████████████▏                 | 622/1230 [4:42:17<4:43:10, 27.94s/it]
                                                                                
{'loss': 0.7052, 'grad_norm': 0.3961451053619385, 'learning_rate': 0.00010155705253094967, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.83, 'epoch': 0.44}

 51%|██████████████████▏                 | 622/1230 [4:42:17<4:43:10, 27.94s/it]
 51%|██████████████████▏                 | 623/1230 [4:42:42<4:34:32, 27.14s/it]
                                                                                
{'loss': 0.7994, 'grad_norm': 0.4195573329925537, 'learning_rate': 0.00010134402702873199, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.75, 'epoch': 0.44}

 51%|██████████████████▏                 | 623/1230 [4:42:42<4:34:32, 27.14s/it]
 51%|██████████████████▎                 | 624/1230 [4:43:12<4:42:17, 27.95s/it]
                                                                                
{'loss': 0.7513, 'grad_norm': 0.4084070324897766, 'learning_rate': 0.0001011309852989456, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.6, 'epoch': 0.44}

 51%|██████████████████▎                 | 624/1230 [4:43:12<4:42:17, 27.95s/it]
 51%|██████████████████▎                 | 625/1230 [4:43:42<4:46:42, 28.43s/it]
                                                                                
{'loss': 0.751, 'grad_norm': 0.3890320062637329, 'learning_rate': 0.00010091792881646654, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.96, 'epoch': 0.44}

 51%|██████████████████▎                 | 625/1230 [4:43:42<4:46:42, 28.43s/it]
 51%|██████████████████▎                 | 626/1230 [4:44:08<4:39:21, 27.75s/it]
                                                                                
{'loss': 0.7398, 'grad_norm': 0.4072780907154083, 'learning_rate': 0.00010070485905627302, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.81, 'epoch': 0.44}

 51%|██████████████████▎                 | 626/1230 [4:44:08<4:39:21, 27.75s/it]
 51%|██████████████████▎                 | 627/1230 [4:44:37<4:43:13, 28.18s/it]
                                                                                
{'loss': 0.7506, 'grad_norm': 0.3773830831050873, 'learning_rate': 0.00010049177749343512, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.04, 'epoch': 0.44}

 51%|██████████████████▎                 | 627/1230 [4:44:37<4:43:13, 28.18s/it]
 51%|██████████████████▍                 | 628/1230 [4:45:03<4:35:33, 27.46s/it]
                                                                                
{'loss': 0.7052, 'grad_norm': 0.40586546063423157, 'learning_rate': 0.00010027868560310461, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.07, 'epoch': 0.44}

 51%|██████████████████▍                 | 628/1230 [4:45:03<4:35:33, 27.46s/it]
 51%|██████████████████▍                 | 629/1230 [4:45:32<4:40:09, 27.97s/it]
                                                                                
{'loss': 0.7497, 'grad_norm': 0.387017160654068, 'learning_rate': 0.00010006558486050482, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.44, 'epoch': 0.45}

 51%|██████████████████▍                 | 629/1230 [4:45:32<4:40:09, 27.97s/it]
 51%|██████████████████▍                 | 630/1230 [4:45:57<4:31:28, 27.15s/it]
                                                                                
{'loss': 0.7581, 'grad_norm': 0.42747393250465393, 'learning_rate': 9.985247674092032e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.68, 'epoch': 0.45}

 51%|██████████████████▍                 | 630/1230 [4:45:57<4:31:28, 27.15s/it]
 51%|██████████████████▍                 | 631/1230 [4:46:23<4:25:44, 26.62s/it]
                                                                                
{'loss': 0.7115, 'grad_norm': 0.40759170055389404, 'learning_rate': 9.963936271968679e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.94, 'epoch': 0.45}

 51%|██████████████████▍                 | 631/1230 [4:46:23<4:25:44, 26.62s/it]
 51%|██████████████████▍                 | 632/1230 [4:46:52<4:34:17, 27.52s/it]
                                                                                
{'loss': 0.728, 'grad_norm': 0.406656414270401, 'learning_rate': 9.942624427218071e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.22, 'epoch': 0.45}

 51%|██████████████████▍                 | 632/1230 [4:46:52<4:34:17, 27.52s/it]
 51%|██████████████████▌                 | 633/1230 [4:47:22<4:39:37, 28.10s/it]
                                                                                
{'loss': 0.7711, 'grad_norm': 0.3956465721130371, 'learning_rate': 9.921312287380927e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.48, 'epoch': 0.45}

 51%|██████████████████▌                 | 633/1230 [4:47:22<4:39:37, 28.10s/it]
 52%|██████████████████▌                 | 634/1230 [4:47:51<4:42:47, 28.47s/it]
                                                                                
{'loss': 0.7371, 'grad_norm': 0.4001268148422241, 'learning_rate': 9.900000000000001e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.9, 'epoch': 0.45}

 52%|██████████████████▌                 | 634/1230 [4:47:51<4:42:47, 28.47s/it]
 52%|██████████████████▌                 | 635/1230 [4:48:20<4:44:48, 28.72s/it]
                                                                                
{'loss': 0.7187, 'grad_norm': 0.3922140300273895, 'learning_rate': 9.878687712619076e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.06, 'epoch': 0.45}

 52%|██████████████████▌                 | 635/1230 [4:48:20<4:44:48, 28.72s/it]
 52%|██████████████████▌                 | 636/1230 [4:48:46<4:34:15, 27.70s/it]
                                                                                
{'loss': 0.753, 'grad_norm': 0.40897610783576965, 'learning_rate': 9.857375572781929e-05, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 207.22, 'epoch': 0.45}

 52%|██████████████████▌                 | 636/1230 [4:48:46<4:34:15, 27.70s/it]
 52%|██████████████████▋                 | 637/1230 [4:49:15<4:39:18, 28.26s/it]
                                                                                
{'loss': 0.7437, 'grad_norm': 0.4038027226924896, 'learning_rate': 9.836063728031324e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.17, 'epoch': 0.45}

 52%|██████████████████▋                 | 637/1230 [4:49:15<4:39:18, 28.26s/it]
 52%|██████████████████▋                 | 638/1230 [4:49:45<4:42:32, 28.64s/it]
                                                                                
{'loss': 0.7557, 'grad_norm': 0.41619667410850525, 'learning_rate': 9.814752325907971e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.36, 'epoch': 0.45}

 52%|██████████████████▋                 | 638/1230 [4:49:45<4:42:32, 28.64s/it]
 52%|██████████████████▋                 | 639/1230 [4:50:10<4:32:13, 27.64s/it]
                                                                                
{'loss': 0.7393, 'grad_norm': 0.4181324243545532, 'learning_rate': 9.793441513949521e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.8, 'epoch': 0.45}

 52%|██████████████████▋                 | 639/1230 [4:50:10<4:32:13, 27.64s/it]
 52%|██████████████████▋                 | 640/1230 [4:50:40<4:37:07, 28.18s/it]
                                                                                
{'loss': 0.7339, 'grad_norm': 0.3913358747959137, 'learning_rate': 9.772131439689542e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.98, 'epoch': 0.45}

 52%|██████████████████▋                 | 640/1230 [4:50:40<4:37:07, 28.18s/it]
 52%|██████████████████▊                 | 641/1230 [4:51:05<4:28:10, 27.32s/it]
                                                                                
{'loss': 0.8085, 'grad_norm': 0.42911267280578613, 'learning_rate': 9.75082225065649e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.73, 'epoch': 0.45}

 52%|██████████████████▊                 | 641/1230 [4:51:05<4:28:10, 27.32s/it]
 52%|██████████████████▊                 | 642/1230 [4:51:31<4:22:57, 26.83s/it]
                                                                                
{'loss': 0.7239, 'grad_norm': 0.3738200068473816, 'learning_rate': 9.729514094372701e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.32, 'epoch': 0.45}

 52%|██████████████████▊                 | 642/1230 [4:51:31<4:22:57, 26.83s/it]
 52%|██████████████████▊                 | 643/1230 [4:52:00<4:29:53, 27.59s/it]
                                                                                
{'loss': 0.7155, 'grad_norm': 0.39388707280158997, 'learning_rate': 9.708207118353349e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.75, 'epoch': 0.46}

 52%|██████████████████▊                 | 643/1230 [4:52:00<4:29:53, 27.59s/it]
 52%|██████████████████▊                 | 644/1230 [4:52:25<4:22:10, 26.84s/it]
                                                                                
{'loss': 0.7295, 'grad_norm': 0.41002151370048523, 'learning_rate': 9.686901470105443e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.55, 'epoch': 0.46}

 52%|██████████████████▊                 | 644/1230 [4:52:25<4:22:10, 26.84s/it]
 52%|██████████████████▉                 | 645/1230 [4:52:50<4:17:00, 26.36s/it]
                                                                                
{'loss': 0.7509, 'grad_norm': 0.40021586418151855, 'learning_rate': 9.665597297126804e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.34, 'epoch': 0.46}

 52%|██████████████████▉                 | 645/1230 [4:52:50<4:17:00, 26.36s/it]
 53%|██████████████████▉                 | 646/1230 [4:53:20<4:25:48, 27.31s/it]
                                                                                
{'loss': 0.7422, 'grad_norm': 0.3899441361427307, 'learning_rate': 9.644294746905036e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.01, 'epoch': 0.46}

 53%|██████████████████▉                 | 646/1230 [4:53:20<4:25:48, 27.31s/it]
 53%|██████████████████▉                 | 647/1230 [4:53:45<4:19:31, 26.71s/it]
                                                                                
{'loss': 0.7166, 'grad_norm': 0.4025310277938843, 'learning_rate': 9.62299396691651e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.04, 'epoch': 0.46}

 53%|██████████████████▉                 | 647/1230 [4:53:45<4:19:31, 26.71s/it]
 53%|██████████████████▉                 | 648/1230 [4:54:10<4:14:28, 26.23s/it]
                                                                                
{'loss': 0.7466, 'grad_norm': 0.39822879433631897, 'learning_rate': 9.601695104625334e-05, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.33, 'epoch': 0.46}

 53%|██████████████████▉                 | 648/1230 [4:54:10<4:14:28, 26.23s/it]
 53%|██████████████████▉                 | 649/1230 [4:54:36<4:13:01, 26.13s/it]
                                                                                
{'loss': 0.7555, 'grad_norm': 0.3863402307033539, 'learning_rate': 9.580398307482354e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.05, 'epoch': 0.46}

 53%|██████████████████▉                 | 649/1230 [4:54:36<4:13:01, 26.13s/it]
 53%|███████████████████                 | 650/1230 [4:55:05<4:21:56, 27.10s/it]
                                                                                
{'loss': 0.7454, 'grad_norm': 0.4006320536136627, 'learning_rate': 9.559103722924108e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.55, 'epoch': 0.46}

 53%|███████████████████                 | 650/1230 [4:55:05<4:21:56, 27.10s/it]
 53%|███████████████████                 | 651/1230 [4:55:31<4:18:06, 26.75s/it]
                                                                                
{'loss': 0.7494, 'grad_norm': 0.3805828392505646, 'learning_rate': 9.537811498371822e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.54, 'epoch': 0.46}

 53%|███████████████████                 | 651/1230 [4:55:31<4:18:06, 26.75s/it]
 53%|███████████████████                 | 652/1230 [4:56:01<4:25:51, 27.60s/it]
                                                                                
{'loss': 0.6707, 'grad_norm': 0.3697086274623871, 'learning_rate': 9.516521781230383e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.38, 'epoch': 0.46}

 53%|███████████████████                 | 652/1230 [4:56:01<4:25:51, 27.60s/it]
 53%|███████████████████                 | 653/1230 [4:56:26<4:18:59, 26.93s/it]
                                                                                
{'loss': 0.6902, 'grad_norm': 0.40695399045944214, 'learning_rate': 9.495234718887312e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.06, 'epoch': 0.46}

 53%|███████████████████                 | 653/1230 [4:56:26<4:18:59, 26.93s/it]
 53%|███████████████████▏                | 654/1230 [4:56:56<4:27:04, 27.82s/it]
                                                                                
{'loss': 0.7469, 'grad_norm': 0.4096183776855469, 'learning_rate': 9.473950458711764e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.24, 'epoch': 0.46}

 53%|███████████████████▏                | 654/1230 [4:56:56<4:27:04, 27.82s/it]
 53%|███████████████████▏                | 655/1230 [4:57:26<4:31:59, 28.38s/it]
                                                                                
{'loss': 0.7535, 'grad_norm': 0.387295663356781, 'learning_rate': 9.452669148053486e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.89, 'epoch': 0.46}

 53%|███████████████████▏                | 655/1230 [4:57:26<4:31:59, 28.38s/it]
 53%|███████████████████▏                | 656/1230 [4:57:52<4:25:11, 27.72s/it]
                                                                                
{'loss': 0.7314, 'grad_norm': 0.39428016543388367, 'learning_rate': 9.431390934241805e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.22, 'epoch': 0.46}

 53%|███████████████████▏                | 656/1230 [4:57:52<4:25:11, 27.72s/it]
 53%|███████████████████▏                | 657/1230 [4:58:18<4:20:07, 27.24s/it]
                                                                                
{'loss': 0.7193, 'grad_norm': 0.36532431840896606, 'learning_rate': 9.410115964584616e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.66, 'epoch': 0.46}

 53%|███████████████████▏                | 657/1230 [4:58:18<4:20:07, 27.24s/it]
 53%|███████████████████▎                | 658/1230 [4:58:44<4:16:40, 26.92s/it]
                                                                                
{'loss': 0.7353, 'grad_norm': 0.41639789938926697, 'learning_rate': 9.388844386367346e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.44, 'epoch': 0.47}

 53%|███████████████████▎                | 658/1230 [4:58:44<4:16:40, 26.92s/it]
 54%|███████████████████▎                | 659/1230 [4:59:13<4:21:50, 27.51s/it]
                                                                                
{'loss': 0.6997, 'grad_norm': 0.36079931259155273, 'learning_rate': 9.367576346851948e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.68, 'epoch': 0.47}

 54%|███████████████████▎                | 659/1230 [4:59:13<4:21:50, 27.51s/it]
 54%|███████████████████▎                | 660/1230 [4:59:38<4:13:58, 26.73s/it]
                                                                                
{'loss': 0.727, 'grad_norm': 0.4217802584171295, 'learning_rate': 9.346311993275878e-05, 'memory/max_active (GiB)': 10.52, 'memory/max_allocated (GiB)': 10.52, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.37, 'epoch': 0.47}

 54%|███████████████████▎                | 660/1230 [4:59:38<4:13:58, 26.73s/it]
 54%|███████████████████▎                | 661/1230 [5:00:03<4:08:26, 26.20s/it]
                                                                                
{'loss': 0.7366, 'grad_norm': 0.41336092352867126, 'learning_rate': 9.325051472851071e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.8, 'epoch': 0.47}

 54%|███████████████████▎                | 661/1230 [5:00:03<4:08:26, 26.20s/it]
 54%|███████████████████▍                | 662/1230 [5:00:32<4:16:19, 27.08s/it]
                                                                                
{'loss': 0.7507, 'grad_norm': 0.4121139645576477, 'learning_rate': 9.303794932762932e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.03, 'epoch': 0.47}

 54%|███████████████████▍                | 662/1230 [5:00:32<4:16:19, 27.08s/it]
 54%|███████████████████▍                | 663/1230 [5:00:58<4:10:43, 26.53s/it]
                                                                                
{'loss': 0.7356, 'grad_norm': 0.4016611576080322, 'learning_rate': 9.282542520169298e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.13, 'epoch': 0.47}

 54%|███████████████████▍                | 663/1230 [5:00:58<4:10:43, 26.53s/it]
 54%|███████████████████▍                | 664/1230 [5:01:27<4:19:21, 27.49s/it]
                                                                                
{'loss': 0.7507, 'grad_norm': 0.396841824054718, 'learning_rate': 9.261294382199441e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.48, 'epoch': 0.47}

 54%|███████████████████▍                | 664/1230 [5:01:27<4:19:21, 27.49s/it]
 54%|███████████████████▍                | 665/1230 [5:01:57<4:23:52, 28.02s/it]
                                                                                
{'loss': 0.7518, 'grad_norm': 0.3906061053276062, 'learning_rate': 9.240050665953041e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.17, 'epoch': 0.47}

 54%|███████████████████▍                | 665/1230 [5:01:57<4:23:52, 28.02s/it]
 54%|███████████████████▍                | 666/1230 [5:02:26<4:27:14, 28.43s/it]
                                                                                
{'loss': 0.7178, 'grad_norm': 0.38233527541160583, 'learning_rate': 9.21881151849916e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.34, 'epoch': 0.47}

 54%|███████████████████▍                | 666/1230 [5:02:26<4:27:14, 28.43s/it]
 54%|███████████████████▌                | 667/1230 [5:02:55<4:29:23, 28.71s/it]
                                                                                
{'loss': 0.7313, 'grad_norm': 0.38109150528907776, 'learning_rate': 9.19757708687524e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.45, 'epoch': 0.47}

 54%|███████████████████▌                | 667/1230 [5:02:55<4:29:23, 28.71s/it]
 54%|███████████████████▌                | 668/1230 [5:03:21<4:20:54, 27.86s/it]
                                                                                
{'loss': 0.7495, 'grad_norm': 0.44236475229263306, 'learning_rate': 9.176347518086062e-05, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.14, 'epoch': 0.47}

 54%|███████████████████▌                | 668/1230 [5:03:21<4:20:54, 27.86s/it]
 54%|███████████████████▌                | 669/1230 [5:03:47<4:13:44, 27.14s/it]
                                                                                
{'loss': 0.7126, 'grad_norm': 0.40982428193092346, 'learning_rate': 9.155122959102755e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.53, 'epoch': 0.47}

 54%|███████████████████▌                | 669/1230 [5:03:47<4:13:44, 27.14s/it]
 54%|███████████████████▌                | 670/1230 [5:04:12<4:08:01, 26.57s/it]
                                                                                
{'loss': 0.6762, 'grad_norm': 0.40296363830566406, 'learning_rate': 9.133903556861758e-05, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.71, 'epoch': 0.47}

 54%|███████████████████▌                | 670/1230 [5:04:12<4:08:01, 26.57s/it]
 55%|███████████████████▋                | 671/1230 [5:04:41<4:14:54, 27.36s/it]
                                                                                
{'loss': 0.7206, 'grad_norm': 0.39922481775283813, 'learning_rate': 9.112689458263809e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.13, 'epoch': 0.47}

 55%|███████████████████▋                | 671/1230 [5:04:41<4:14:54, 27.36s/it]
 55%|███████████████████▋                | 672/1230 [5:05:10<4:19:32, 27.91s/it]
                                                                                
{'loss': 0.7598, 'grad_norm': 0.4172070324420929, 'learning_rate': 9.091480810172939e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.72, 'epoch': 0.48}

 55%|███████████████████▋                | 672/1230 [5:05:10<4:19:32, 27.91s/it]
 55%|███████████████████▋                | 673/1230 [5:05:35<4:11:26, 27.09s/it]
                                                                                
{'loss': 0.783, 'grad_norm': 0.458820104598999, 'learning_rate': 9.070277759415435e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.17, 'epoch': 0.48}

 55%|███████████████████▋                | 673/1230 [5:05:35<4:11:26, 27.09s/it]
 55%|███████████████████▋                | 674/1230 [5:06:05<4:16:48, 27.71s/it]
                                                                                
{'loss': 0.7364, 'grad_norm': 0.39011287689208984, 'learning_rate': 9.049080452778835e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.74, 'epoch': 0.48}

 55%|███████████████████▋                | 674/1230 [5:06:05<4:16:48, 27.71s/it]
 55%|███████████████████▊                | 675/1230 [5:06:31<4:12:09, 27.26s/it]
                                                                                
{'loss': 0.7902, 'grad_norm': 0.4089367091655731, 'learning_rate': 9.027889037010918e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.21, 'epoch': 0.48}

 55%|███████████████████▊                | 675/1230 [5:06:31<4:12:09, 27.26s/it]
 55%|███████████████████▊                | 676/1230 [5:07:00<4:18:32, 28.00s/it]
                                                                                
{'loss': 0.77, 'grad_norm': 0.4175400733947754, 'learning_rate': 9.006703658818672e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.35, 'epoch': 0.48}

 55%|███████████████████▊                | 676/1230 [5:07:01<4:18:32, 28.00s/it]
 55%|███████████████████▊                | 677/1230 [5:07:26<4:10:48, 27.21s/it]
                                                                                
{'loss': 0.7415, 'grad_norm': 0.4283519685268402, 'learning_rate': 8.9855244648673e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.98, 'epoch': 0.48}

 55%|███████████████████▊                | 677/1230 [5:07:26<4:10:48, 27.21s/it]
 55%|███████████████████▊                | 678/1230 [5:07:55<4:16:51, 27.92s/it]
                                                                                
{'loss': 0.7562, 'grad_norm': 0.398315966129303, 'learning_rate': 8.964351601779172e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.06, 'epoch': 0.48}

 55%|███████████████████▊                | 678/1230 [5:07:55<4:16:51, 27.92s/it]
 55%|███████████████████▊                | 679/1230 [5:08:21<4:09:31, 27.17s/it]
                                                                                
{'loss': 0.7425, 'grad_norm': 0.4053307771682739, 'learning_rate': 8.943185216132846e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.19, 'epoch': 0.48}

 55%|███████████████████▊                | 679/1230 [5:08:21<4:09:31, 27.17s/it]
 55%|███████████████████▉                | 680/1230 [5:08:51<4:16:16, 27.96s/it]
                                                                                
{'loss': 0.7408, 'grad_norm': 0.3832851052284241, 'learning_rate': 8.922025454462036e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.62, 'epoch': 0.48}

 55%|███████████████████▉                | 680/1230 [5:08:51<4:16:16, 27.96s/it]
 55%|███████████████████▉                | 681/1230 [5:09:16<4:08:06, 27.12s/it]
                                                                                
{'loss': 0.7622, 'grad_norm': 0.415097177028656, 'learning_rate': 8.900872463254587e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.82, 'epoch': 0.48}

 55%|███████████████████▉                | 681/1230 [5:09:16<4:08:06, 27.12s/it]
 55%|███████████████████▉                | 682/1230 [5:09:41<4:01:54, 26.49s/it]
                                                                                
{'loss': 0.6891, 'grad_norm': 0.41470783948898315, 'learning_rate': 8.879726388951489e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.27, 'epoch': 0.48}

 55%|███████████████████▉                | 682/1230 [5:09:41<4:01:54, 26.49s/it]
 56%|███████████████████▉                | 683/1230 [5:10:07<3:59:29, 26.27s/it]
                                                                                
{'loss': 0.7232, 'grad_norm': 0.4079004228115082, 'learning_rate': 8.858587377945831e-05, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.48, 'epoch': 0.48}

 56%|███████████████████▉                | 683/1230 [5:10:07<3:59:29, 26.27s/it]
 56%|████████████████████                | 684/1230 [5:10:36<4:06:53, 27.13s/it]
                                                                                
{'loss': 0.7072, 'grad_norm': 0.39589717984199524, 'learning_rate': 8.837455576581815e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.92, 'epoch': 0.48}

 56%|████████████████████                | 684/1230 [5:10:36<4:06:53, 27.13s/it]
 56%|████████████████████                | 685/1230 [5:11:01<4:02:05, 26.65s/it]
                                                                                
{'loss': 0.7077, 'grad_norm': 0.3835257887840271, 'learning_rate': 8.816331131153721e-05, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 210.1, 'epoch': 0.48}

 56%|████████████████████                | 685/1230 [5:11:01<4:02:05, 26.65s/it]
 56%|████████████████████                | 686/1230 [5:11:28<4:02:24, 26.74s/it]
                                                                                
{'loss': 0.753, 'grad_norm': 0.4203863739967346, 'learning_rate': 8.795214187904913e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.19, 'epoch': 0.49}

 56%|████████████████████                | 686/1230 [5:11:28<4:02:24, 26.74s/it]
 56%|████████████████████                | 687/1230 [5:11:54<3:59:43, 26.49s/it]
                                                                                
{'loss': 0.7528, 'grad_norm': 0.4146800637245178, 'learning_rate': 8.774104893026813e-05, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.12, 'epoch': 0.49}

 56%|████████████████████                | 687/1230 [5:11:54<3:59:43, 26.49s/it]
 56%|████████████████████▏               | 688/1230 [5:12:23<4:07:07, 27.36s/it]
                                                                                
{'loss': 0.7119, 'grad_norm': 0.4130697250366211, 'learning_rate': 8.753003392657895e-05, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 142.44, 'epoch': 0.49}

 56%|████████████████████▏               | 688/1230 [5:12:24<4:07:07, 27.36s/it]
 56%|████████████████████▏               | 689/1230 [5:12:53<4:13:13, 28.08s/it]
                                                                                
{'loss': 0.7244, 'grad_norm': 0.39328432083129883, 'learning_rate': 8.731909832882671e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.84, 'epoch': 0.49}

 56%|████████████████████▏               | 689/1230 [5:12:53<4:13:13, 28.08s/it]
 56%|████████████████████▏               | 690/1230 [5:13:19<4:05:37, 27.29s/it]
                                                                                
{'loss': 0.6864, 'grad_norm': 0.41772809624671936, 'learning_rate': 8.710824359730682e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.9, 'epoch': 0.49}

 56%|████████████████████▏               | 690/1230 [5:13:19<4:05:37, 27.29s/it]
 56%|████████████████████▏               | 691/1230 [5:13:48<4:11:30, 28.00s/it]
                                                                                
{'loss': 0.7769, 'grad_norm': 0.41056495904922485, 'learning_rate': 8.689747119175488e-05, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.96, 'epoch': 0.49}

 56%|████████████████████▏               | 691/1230 [5:13:48<4:11:30, 28.00s/it]
 56%|████████████████████▎               | 692/1230 [5:14:15<4:06:04, 27.44s/it]
                                                                                
{'loss': 0.6958, 'grad_norm': 0.3989245593547821, 'learning_rate': 8.668678257133651e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.03, 'epoch': 0.49}

 56%|████████████████████▎               | 692/1230 [5:14:15<4:06:04, 27.44s/it]
 56%|████████████████████▎               | 693/1230 [5:14:41<4:02:51, 27.13s/it]
                                                                                
{'loss': 0.6849, 'grad_norm': 0.38483288884162903, 'learning_rate': 8.64761791946373e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.32, 'epoch': 0.49}

 56%|████████████████████▎               | 693/1230 [5:14:41<4:02:51, 27.13s/it]
 56%|████████████████████▎               | 694/1230 [5:15:06<3:56:18, 26.45s/it]
                                                                                
{'loss': 0.7613, 'grad_norm': 0.4191681742668152, 'learning_rate': 8.626566251965269e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.26, 'epoch': 0.49}

 56%|████████████████████▎               | 694/1230 [5:15:06<3:56:18, 26.45s/it]
 57%|████████████████████▎               | 695/1230 [5:15:31<3:51:22, 25.95s/it]
                                                                                
{'loss': 0.7637, 'grad_norm': 0.42681631445884705, 'learning_rate': 8.605523400377796e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.99, 'epoch': 0.49}

 57%|████████████████████▎               | 695/1230 [5:15:31<3:51:22, 25.95s/it]
 57%|████████████████████▎               | 696/1230 [5:16:00<3:59:51, 26.95s/it]
                                                                                
{'loss': 0.7382, 'grad_norm': 0.37707746028900146, 'learning_rate': 8.5844895103798e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.52, 'epoch': 0.49}

 57%|████████████████████▎               | 696/1230 [5:16:00<3:59:51, 26.95s/it]
 57%|████████████████████▍               | 697/1230 [5:16:25<3:55:01, 26.46s/it]
                                                                                
{'loss': 0.7688, 'grad_norm': 0.40127602219581604, 'learning_rate': 8.563464727587728e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.77, 'epoch': 0.49}

 57%|████████████████████▍               | 697/1230 [5:16:25<3:55:01, 26.46s/it]
 57%|████████████████████▍               | 698/1230 [5:16:55<4:02:57, 27.40s/it]
                                                                                
{'loss': 0.7316, 'grad_norm': 0.3923484683036804, 'learning_rate': 8.542449197554984e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.54, 'epoch': 0.49}

 57%|████████████████████▍               | 698/1230 [5:16:55<4:02:57, 27.40s/it]
 57%|████████████████████▍               | 699/1230 [5:17:24<4:08:19, 28.06s/it]
                                                                                
{'loss': 0.7637, 'grad_norm': 0.39591845870018005, 'learning_rate': 8.521443065770912e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.16, 'epoch': 0.49}

 57%|████████████████████▍               | 699/1230 [5:17:24<4:08:19, 28.06s/it]
 57%|████████████████████▍               | 700/1230 [5:17:54<4:11:53, 28.52s/it]
                                                                                
{'loss': 0.7389, 'grad_norm': 0.41113749146461487, 'learning_rate': 8.500446477659796e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.06, 'epoch': 0.5}

 57%|████████████████████▍               | 700/1230 [5:17:54<4:11:53, 28.52s/it]
 57%|████████████████████▌               | 701/1230 [5:18:19<4:02:40, 27.52s/it]
                                                                                
{'loss': 0.7329, 'grad_norm': 0.41112157702445984, 'learning_rate': 8.479459578579842e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.54, 'epoch': 0.5}

 57%|████████████████████▌               | 701/1230 [5:18:19<4:02:40, 27.52s/it]
 57%|████████████████████▌               | 702/1230 [5:18:49<4:07:11, 28.09s/it]
                                                                                
{'loss': 0.6843, 'grad_norm': 0.382368803024292, 'learning_rate': 8.458482513822194e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.52, 'epoch': 0.5}

 57%|████████████████████▌               | 702/1230 [5:18:49<4:07:11, 28.09s/it]
 57%|████████████████████▌               | 703/1230 [5:19:14<3:58:45, 27.18s/it]
                                                                                
{'loss': 0.745, 'grad_norm': 0.404814213514328, 'learning_rate': 8.437515428609894e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.83, 'epoch': 0.5}

 57%|████████████████████▌               | 703/1230 [5:19:14<3:58:45, 27.18s/it]
 57%|████████████████████▌               | 704/1230 [5:19:43<4:03:57, 27.83s/it]
                                                                                
{'loss': 0.7241, 'grad_norm': 0.41305989027023315, 'learning_rate': 8.416558468096913e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.93, 'epoch': 0.5}

 57%|████████████████████▌               | 704/1230 [5:19:43<4:03:57, 27.83s/it]
 57%|████████████████████▋               | 705/1230 [5:20:12<4:07:06, 28.24s/it]
                                                                                
{'loss': 0.7161, 'grad_norm': 0.4044142961502075, 'learning_rate': 8.395611777367118e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.3, 'epoch': 0.5}

 57%|████████████████████▋               | 705/1230 [5:20:12<4:07:06, 28.24s/it]
 57%|████████████████████▋               | 706/1230 [5:20:38<3:59:59, 27.48s/it]
                                                                                
{'loss': 0.7025, 'grad_norm': 0.3858373165130615, 'learning_rate': 8.374675501433285e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.64, 'epoch': 0.5}

 57%|████████████████████▋               | 706/1230 [5:20:38<3:59:59, 27.48s/it]
 57%|████████████████████▋               | 707/1230 [5:21:07<4:04:15, 28.02s/it]
                                                                                
{'loss': 0.6975, 'grad_norm': 0.39531829953193665, 'learning_rate': 8.353749785236092e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.85, 'epoch': 0.5}

 57%|████████████████████▋               | 707/1230 [5:21:07<4:04:15, 28.02s/it]
 58%|████████████████████▋               | 708/1230 [5:21:33<3:59:00, 27.47s/it]
                                                                                
{'loss': 0.7224, 'grad_norm': 0.41106313467025757, 'learning_rate': 8.332834773643103e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.1, 'epoch': 0.5}

 58%|████████████████████▋               | 708/1230 [5:21:33<3:59:00, 27.47s/it]
 58%|████████████████████▊               | 709/1230 [5:22:03<4:03:13, 28.01s/it]
                                                                                
{'loss': 0.7812, 'grad_norm': 0.4214148223400116, 'learning_rate': 8.31193061144778e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.42, 'epoch': 0.5}

 58%|████████████████████▊               | 709/1230 [5:22:03<4:03:13, 28.01s/it]
 58%|████████████████████▊               | 710/1230 [5:22:28<3:55:37, 27.19s/it]
                                                                                
{'loss': 0.7193, 'grad_norm': 0.41732335090637207, 'learning_rate': 8.291037443368473e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.7, 'epoch': 0.5}

 58%|████████████████████▊               | 710/1230 [5:22:28<3:55:37, 27.19s/it]
 58%|████████████████████▊               | 711/1230 [5:22:53<3:50:32, 26.65s/it]
                                                                                
{'loss': 0.7191, 'grad_norm': 0.3993341326713562, 'learning_rate': 8.270155414047428e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.99, 'epoch': 0.5}

 58%|████████████████████▊               | 711/1230 [5:22:53<3:50:32, 26.65s/it]
 58%|████████████████████▊               | 712/1230 [5:23:19<3:48:08, 26.43s/it]
                                                                                
{'loss': 0.6918, 'grad_norm': 0.37492692470550537, 'learning_rate': 8.24928466804977e-05, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.77, 'epoch': 0.5}

 58%|████████████████████▊               | 712/1230 [5:23:19<3:48:08, 26.43s/it]
 58%|████████████████████▊               | 713/1230 [5:23:49<3:55:17, 27.31s/it]
                                                                                
{'loss': 0.7247, 'grad_norm': 0.3919753134250641, 'learning_rate': 8.228425349862509e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.81, 'epoch': 0.5}

 58%|████████████████████▊               | 713/1230 [5:23:49<3:55:17, 27.31s/it]
 58%|████████████████████▉               | 714/1230 [5:24:18<4:00:43, 27.99s/it]
                                                                                
{'loss': 0.7023, 'grad_norm': 0.4225118160247803, 'learning_rate': 8.207577603893548e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.52, 'epoch': 0.51}

 58%|████████████████████▉               | 714/1230 [5:24:18<4:00:43, 27.99s/it]
 58%|████████████████████▉               | 715/1230 [5:24:48<4:04:07, 28.44s/it]
                                                                                
{'loss': 0.6933, 'grad_norm': 0.3887205719947815, 'learning_rate': 8.186741574470667e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.06, 'epoch': 0.51}

 58%|████████████████████▉               | 715/1230 [5:24:48<4:04:07, 28.44s/it]
 58%|████████████████████▉               | 716/1230 [5:25:13<3:56:16, 27.58s/it]
                                                                                
{'loss': 0.7033, 'grad_norm': 0.3901570439338684, 'learning_rate': 8.165917405840542e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.42, 'epoch': 0.51}

 58%|████████████████████▉               | 716/1230 [5:25:13<3:56:16, 27.58s/it]
 58%|████████████████████▉               | 717/1230 [5:25:38<3:49:21, 26.83s/it]
                                                                                
{'loss': 0.7048, 'grad_norm': 0.41432225704193115, 'learning_rate': 8.145105242167741e-05, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.67, 'epoch': 0.51}

 58%|████████████████████▉               | 717/1230 [5:25:38<3:49:21, 26.83s/it]
 58%|█████████████████████               | 718/1230 [5:26:08<3:55:20, 27.58s/it]
                                                                                
{'loss': 0.729, 'grad_norm': 0.4119988679885864, 'learning_rate': 8.124305227533703e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.49, 'epoch': 0.51}

 58%|█████████████████████               | 718/1230 [5:26:08<3:55:20, 27.58s/it]
 58%|█████████████████████               | 719/1230 [5:26:37<3:58:58, 28.06s/it]
                                                                                
{'loss': 0.7143, 'grad_norm': 0.4321652352809906, 'learning_rate': 8.103517505935776e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.04, 'epoch': 0.51}

 58%|█████████████████████               | 719/1230 [5:26:37<3:58:58, 28.06s/it]
 59%|█████████████████████               | 720/1230 [5:27:06<4:00:34, 28.30s/it]
                                                                                
{'loss': 0.7255, 'grad_norm': 0.42264634370803833, 'learning_rate': 8.082742221286202e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.09, 'epoch': 0.51}

 59%|█████████████████████               | 720/1230 [5:27:06<4:00:34, 28.30s/it]
 59%|█████████████████████               | 721/1230 [5:27:35<4:02:59, 28.64s/it]
                                                                                
{'loss': 0.7029, 'grad_norm': 0.40423285961151123, 'learning_rate': 8.061979517411119e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.6, 'epoch': 0.51}

 59%|█████████████████████               | 721/1230 [5:27:35<4:02:59, 28.64s/it]
 59%|█████████████████████▏              | 722/1230 [5:28:01<3:55:03, 27.76s/it]
                                                                                
{'loss': 0.7103, 'grad_norm': 0.40173518657684326, 'learning_rate': 8.041229538049577e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.79, 'epoch': 0.51}

 59%|█████████████████████▏              | 722/1230 [5:28:01<3:55:03, 27.76s/it]
 59%|█████████████████████▏              | 723/1230 [5:28:30<3:58:40, 28.25s/it]
                                                                                
{'loss': 0.7469, 'grad_norm': 0.4121905267238617, 'learning_rate': 8.020492426852518e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.97, 'epoch': 0.51}

 59%|█████████████████████▏              | 723/1230 [5:28:30<3:58:40, 28.25s/it]
 59%|█████████████████████▏              | 724/1230 [5:28:56<3:51:01, 27.39s/it]
                                                                                
{'loss': 0.718, 'grad_norm': 0.3956618905067444, 'learning_rate': 7.999768327381814e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.03, 'epoch': 0.51}

 59%|█████████████████████▏              | 724/1230 [5:28:56<3:51:01, 27.39s/it]
 59%|█████████████████████▏              | 725/1230 [5:29:21<3:45:08, 26.75s/it]
                                                                                
{'loss': 0.6784, 'grad_norm': 0.4122350513935089, 'learning_rate': 7.979057383109254e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.24, 'epoch': 0.51}

 59%|█████████████████████▏              | 725/1230 [5:29:21<3:45:08, 26.75s/it]
 59%|█████████████████████▏              | 726/1230 [5:29:46<3:40:43, 26.28s/it]
                                                                                
{'loss': 0.7182, 'grad_norm': 0.38433903455734253, 'learning_rate': 7.958359737415553e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.22, 'epoch': 0.51}

 59%|█████████████████████▏              | 726/1230 [5:29:46<3:40:43, 26.28s/it]
 59%|█████████████████████▎              | 727/1230 [5:30:12<3:38:29, 26.06s/it]
                                                                                
{'loss': 0.7064, 'grad_norm': 0.3885841965675354, 'learning_rate': 7.937675533589358e-05, 'memory/max_active (GiB)': 8.86, 'memory/max_allocated (GiB)': 8.86, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 210.24, 'epoch': 0.51}

 59%|█████████████████████▎              | 727/1230 [5:30:12<3:38:29, 26.06s/it]
 59%|█████████████████████▎              | 728/1230 [5:30:41<3:46:25, 27.06s/it]
                                                                                
{'loss': 0.6995, 'grad_norm': 0.3893852233886719, 'learning_rate': 7.917004914826264e-05, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.15, 'epoch': 0.52}

 59%|█████████████████████▎              | 728/1230 [5:30:41<3:46:25, 27.06s/it]
 59%|█████████████████████▎              | 729/1230 [5:31:10<3:51:29, 27.72s/it]
                                                                                
{'loss': 0.7151, 'grad_norm': 0.3734019994735718, 'learning_rate': 7.896348024227811e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.95, 'epoch': 0.52}

 59%|█████████████████████▎              | 729/1230 [5:31:10<3:51:29, 27.72s/it]
 59%|█████████████████████▎              | 730/1230 [5:31:39<3:54:48, 28.18s/it]
                                                                                
{'loss': 0.7663, 'grad_norm': 0.43244469165802, 'learning_rate': 7.875705004800502e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.3, 'epoch': 0.52}

 59%|█████████████████████▎              | 730/1230 [5:31:39<3:54:48, 28.18s/it]
 59%|█████████████████████▍              | 731/1230 [5:32:09<3:56:45, 28.47s/it]
                                                                                
{'loss': 0.706, 'grad_norm': 0.3909871280193329, 'learning_rate': 7.855075999454821e-05, 'memory/max_active (GiB)': 10.0, 'memory/max_allocated (GiB)': 10.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.06, 'epoch': 0.52}

 59%|█████████████████████▍              | 731/1230 [5:32:09<3:56:45, 28.47s/it]
 60%|█████████████████████▍              | 732/1230 [5:32:38<3:58:35, 28.75s/it]
                                                                                
{'loss': 0.7317, 'grad_norm': 0.4090435206890106, 'learning_rate': 7.834461151004217e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.5, 'epoch': 0.52}

 60%|█████████████████████▍              | 732/1230 [5:32:38<3:58:35, 28.75s/it]
 60%|█████████████████████▍              | 733/1230 [5:33:03<3:49:33, 27.71s/it]
                                                                                
{'loss': 0.7213, 'grad_norm': 0.41325893998146057, 'learning_rate': 7.81386060216414e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.33, 'epoch': 0.52}

 60%|█████████████████████▍              | 733/1230 [5:33:03<3:49:33, 27.71s/it]
 60%|█████████████████████▍              | 734/1230 [5:33:29<3:44:48, 27.19s/it]
                                                                                
{'loss': 0.7302, 'grad_norm': 0.39739319682121277, 'learning_rate': 7.793274495551042e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.26, 'epoch': 0.52}

 60%|█████████████████████▍              | 734/1230 [5:33:29<3:44:48, 27.19s/it]
 60%|█████████████████████▌              | 735/1230 [5:33:59<3:49:57, 27.87s/it]
                                                                                
{'loss': 0.7502, 'grad_norm': 0.3951013386249542, 'learning_rate': 7.772702973681395e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.78, 'epoch': 0.52}

 60%|█████████████████████▌              | 735/1230 [5:33:59<3:49:57, 27.87s/it]
 60%|█████████████████████▌              | 736/1230 [5:34:28<3:52:17, 28.21s/it]
                                                                                
{'loss': 0.7372, 'grad_norm': 0.41152817010879517, 'learning_rate': 7.752146178970702e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.87, 'epoch': 0.52}

 60%|█████████████████████▌              | 736/1230 [5:34:28<3:52:17, 28.21s/it]
 60%|█████████████████████▌              | 737/1230 [5:34:53<3:44:02, 27.27s/it]
                                                                                
{'loss': 0.733, 'grad_norm': 0.38812094926834106, 'learning_rate': 7.731604253732512e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.17, 'epoch': 0.52}

 60%|█████████████████████▌              | 737/1230 [5:34:53<3:44:02, 27.27s/it]
 60%|█████████████████████▌              | 738/1230 [5:35:18<3:37:50, 26.57s/it]
                                                                                
{'loss': 0.7179, 'grad_norm': 0.4074961245059967, 'learning_rate': 7.711077340177422e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.04, 'epoch': 0.52}

 60%|█████████████████████▌              | 738/1230 [5:35:18<3:37:50, 26.57s/it]
 60%|█████████████████████▋              | 739/1230 [5:35:44<3:36:20, 26.44s/it]
                                                                                
{'loss': 0.7257, 'grad_norm': 0.4048783779144287, 'learning_rate': 7.690565580412116e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.51, 'epoch': 0.52}

 60%|█████████████████████▋              | 739/1230 [5:35:44<3:36:20, 26.44s/it]
 60%|█████████████████████▋              | 740/1230 [5:36:10<3:35:09, 26.35s/it]
                                                                                
{'loss': 0.7124, 'grad_norm': 0.415553480386734, 'learning_rate': 7.67006911643837e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.25, 'epoch': 0.52}

 60%|█████████████████████▋              | 740/1230 [5:36:10<3:35:09, 26.35s/it]
 60%|█████████████████████▋              | 741/1230 [5:36:35<3:31:42, 25.98s/it]
                                                                                
{'loss': 0.732, 'grad_norm': 0.4016774594783783, 'learning_rate': 7.649588090152054e-05, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.33, 'epoch': 0.52}

 60%|█████████████████████▋              | 741/1230 [5:36:35<3:31:42, 25.98s/it]
 60%|█████████████████████▋              | 742/1230 [5:37:05<3:40:18, 27.09s/it]
                                                                                
{'loss': 0.7366, 'grad_norm': 0.4173785150051117, 'learning_rate': 7.629122643342185e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.2, 'epoch': 0.53}

 60%|█████████████████████▋              | 742/1230 [5:37:05<3:40:18, 27.09s/it]
 60%|█████████████████████▋              | 743/1230 [5:37:30<3:35:11, 26.51s/it]
                                                                                
{'loss': 0.7417, 'grad_norm': 0.41865622997283936, 'learning_rate': 7.608672917689905e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.48, 'epoch': 0.53}

 60%|█████████████████████▋              | 743/1230 [5:37:30<3:35:11, 26.51s/it]
 60%|█████████████████████▊              | 744/1230 [5:37:59<3:41:47, 27.38s/it]
                                                                                
{'loss': 0.7083, 'grad_norm': 0.40306171774864197, 'learning_rate': 7.58823905476753e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.32, 'epoch': 0.53}

 60%|█████████████████████▊              | 744/1230 [5:37:59<3:41:47, 27.38s/it]
 61%|█████████████████████▊              | 745/1230 [5:38:28<3:45:14, 27.86s/it]
                                                                                
{'loss': 0.7303, 'grad_norm': 0.3933418393135071, 'learning_rate': 7.567821196037554e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.94, 'epoch': 0.53}

 61%|█████████████████████▊              | 745/1230 [5:38:28<3:45:14, 27.86s/it]
 61%|█████████████████████▊              | 746/1230 [5:38:54<3:38:29, 27.09s/it]
                                                                                
{'loss': 0.7153, 'grad_norm': 0.43797293305397034, 'learning_rate': 7.547419482851677e-05, 'memory/max_active (GiB)': 11.04, 'memory/max_allocated (GiB)': 11.04, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.43, 'epoch': 0.53}

 61%|█████████████████████▊              | 746/1230 [5:38:54<3:38:29, 27.09s/it]
 61%|█████████████████████▊              | 747/1230 [5:39:23<3:43:58, 27.82s/it]
                                                                                
{'loss': 0.7195, 'grad_norm': 0.40037330985069275, 'learning_rate': 7.527034056449829e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.85, 'epoch': 0.53}

 61%|█████████████████████▊              | 747/1230 [5:39:23<3:43:58, 27.82s/it]
 61%|█████████████████████▉              | 748/1230 [5:39:49<3:39:47, 27.36s/it]
                                                                                
{'loss': 0.6782, 'grad_norm': 0.4127013385295868, 'learning_rate': 7.506665057959175e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.81, 'epoch': 0.53}

 61%|█████████████████████▉              | 748/1230 [5:39:49<3:39:47, 27.36s/it]
 61%|█████████████████████▉              | 749/1230 [5:40:19<3:44:49, 28.04s/it]
                                                                                
{'loss': 0.7307, 'grad_norm': 0.41181087493896484, 'learning_rate': 7.486312628393163e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.01, 'epoch': 0.53}

 61%|█████████████████████▉              | 749/1230 [5:40:19<3:44:49, 28.04s/it]
 61%|█████████████████████▉              | 750/1230 [5:40:46<3:40:45, 27.60s/it]
                                                                                
{'loss': 0.7202, 'grad_norm': 0.4174041450023651, 'learning_rate': 7.46597690865053e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.81, 'epoch': 0.53}

 61%|█████████████████████▉              | 750/1230 [5:40:46<3:40:45, 27.60s/it]
 61%|█████████████████████▉              | 751/1230 [5:41:14<3:43:09, 27.95s/it]
                                                                                
{'loss': 0.7228, 'grad_norm': 0.4094357192516327, 'learning_rate': 7.44565803951433e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.42, 'epoch': 0.53}

 61%|█████████████████████▉              | 751/1230 [5:41:14<3:43:09, 27.95s/it]
 61%|██████████████████████              | 752/1230 [5:41:43<3:44:40, 28.20s/it]
                                                                                
{'loss': 0.7006, 'grad_norm': 0.4011639952659607, 'learning_rate': 7.425356161650972e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.19, 'epoch': 0.53}

 61%|██████████████████████              | 752/1230 [5:41:43<3:44:40, 28.20s/it]
 61%|██████████████████████              | 753/1230 [5:42:13<3:46:53, 28.54s/it]
                                                                                
{'loss': 0.7158, 'grad_norm': 0.4015175998210907, 'learning_rate': 7.405071415609216e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.71, 'epoch': 0.53}

 61%|██████████████████████              | 753/1230 [5:42:13<3:46:53, 28.54s/it]
 61%|██████████████████████              | 754/1230 [5:42:39<3:40:40, 27.82s/it]
                                                                                
{'loss': 0.69, 'grad_norm': 0.3975571393966675, 'learning_rate': 7.384803941819233e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.92, 'epoch': 0.53}

 61%|██████████████████████              | 754/1230 [5:42:39<3:40:40, 27.82s/it]
 61%|██████████████████████              | 755/1230 [5:43:04<3:34:33, 27.10s/it]
                                                                                
{'loss': 0.7028, 'grad_norm': 0.3871065378189087, 'learning_rate': 7.364553880591622e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.32, 'epoch': 0.53}

 61%|██████████████████████              | 755/1230 [5:43:04<3:34:33, 27.10s/it]
 61%|██████████████████████▏             | 756/1230 [5:43:30<3:30:28, 26.64s/it]
                                                                                
{'loss': 0.7073, 'grad_norm': 0.3907775282859802, 'learning_rate': 7.344321372116424e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.9, 'epoch': 0.54}

 61%|██████████████████████▏             | 756/1230 [5:43:30<3:30:28, 26.64s/it]
 62%|██████████████████████▏             | 757/1230 [5:43:59<3:37:07, 27.54s/it]
                                                                                
{'loss': 0.7214, 'grad_norm': 0.4089638888835907, 'learning_rate': 7.324106556462171e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.34, 'epoch': 0.54}

 62%|██████████████████████▏             | 757/1230 [5:43:59<3:37:07, 27.54s/it]
 62%|██████████████████████▏             | 758/1230 [5:44:29<3:41:45, 28.19s/it]
                                                                                
{'loss': 0.7369, 'grad_norm': 0.3908684253692627, 'learning_rate': 7.303909573574911e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.01, 'epoch': 0.54}

 62%|██████████████████████▏             | 758/1230 [5:44:29<3:41:45, 28.19s/it]
 62%|██████████████████████▏             | 759/1230 [5:44:54<3:34:31, 27.33s/it]
                                                                                
{'loss': 0.7043, 'grad_norm': 0.38865146040916443, 'learning_rate': 7.283730563277227e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.21, 'epoch': 0.54}

 62%|██████████████████████▏             | 759/1230 [5:44:54<3:34:31, 27.33s/it]
 62%|██████████████████████▏             | 760/1230 [5:45:24<3:38:50, 27.94s/it]
                                                                                
{'loss': 0.7106, 'grad_norm': 0.40098243951797485, 'learning_rate': 7.263569665267286e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.69, 'epoch': 0.54}

 62%|██████████████████████▏             | 760/1230 [5:45:24<3:38:50, 27.94s/it]
 62%|██████████████████████▎             | 761/1230 [5:45:53<3:41:38, 28.36s/it]
                                                                                
{'loss': 0.6867, 'grad_norm': 0.4196755588054657, 'learning_rate': 7.243427019117864e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.41, 'epoch': 0.54}

 62%|██████████████████████▎             | 761/1230 [5:45:53<3:41:38, 28.36s/it]
 62%|██████████████████████▎             | 762/1230 [5:46:18<3:33:17, 27.34s/it]
                                                                                
{'loss': 0.7144, 'grad_norm': 0.3907400369644165, 'learning_rate': 7.223302764275386e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.77, 'epoch': 0.54}

 62%|██████████████████████▎             | 762/1230 [5:46:18<3:33:17, 27.34s/it]
 62%|██████████████████████▎             | 763/1230 [5:46:47<3:36:54, 27.87s/it]
                                                                                
{'loss': 0.7349, 'grad_norm': 0.4185091555118561, 'learning_rate': 7.203197040058938e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.84, 'epoch': 0.54}

 62%|██████████████████████▎             | 763/1230 [5:46:47<3:36:54, 27.87s/it]
 62%|██████████████████████▎             | 764/1230 [5:47:12<3:29:40, 27.00s/it]
                                                                                
{'loss': 0.73, 'grad_norm': 0.42346471548080444, 'learning_rate': 7.183109985659333e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.19, 'epoch': 0.54}

 62%|██████████████████████▎             | 764/1230 [5:47:12<3:29:40, 27.00s/it]
 62%|██████████████████████▍             | 765/1230 [5:47:38<3:26:57, 26.70s/it]
                                                                                
{'loss': 0.7509, 'grad_norm': 0.41919007897377014, 'learning_rate': 7.163041740138134e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.71, 'epoch': 0.54}

 62%|██████████████████████▍             | 765/1230 [5:47:38<3:26:57, 26.70s/it]
 62%|██████████████████████▍             | 766/1230 [5:48:07<3:32:27, 27.47s/it]
                                                                                
{'loss': 0.6613, 'grad_norm': 0.3918885290622711, 'learning_rate': 7.142992442426685e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.73, 'epoch': 0.54}

 62%|██████████████████████▍             | 766/1230 [5:48:07<3:32:27, 27.47s/it]
 62%|██████████████████████▍             | 767/1230 [5:48:37<3:36:10, 28.01s/it]
                                                                                
{'loss': 0.7011, 'grad_norm': 0.3968794047832489, 'learning_rate': 7.122962231325165e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.33, 'epoch': 0.54}

 62%|██████████████████████▍             | 767/1230 [5:48:37<3:36:10, 28.01s/it]
 62%|██████████████████████▍             | 768/1230 [5:49:02<3:30:05, 27.28s/it]
                                                                                
{'loss': 0.7396, 'grad_norm': 0.4401998519897461, 'learning_rate': 7.102951245501604e-05, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.86, 'epoch': 0.54}

 62%|██████████████████████▍             | 768/1230 [5:49:02<3:30:05, 27.28s/it]
 63%|██████████████████████▌             | 769/1230 [5:49:27<3:24:48, 26.66s/it]
                                                                                
{'loss': 0.7145, 'grad_norm': 0.41589370369911194, 'learning_rate': 7.082959623490944e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.6, 'epoch': 0.54}

 63%|██████████████████████▌             | 769/1230 [5:49:27<3:24:48, 26.66s/it]
 63%|██████████████████████▌             | 770/1230 [5:49:53<3:20:51, 26.20s/it]
                                                                                
{'loss': 0.6742, 'grad_norm': 0.4088842272758484, 'learning_rate': 7.062987503694075e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.43, 'epoch': 0.54}

 63%|██████████████████████▌             | 770/1230 [5:49:53<3:20:51, 26.20s/it]
 63%|██████████████████████▌             | 771/1230 [5:50:22<3:27:21, 27.11s/it]
                                                                                
{'loss': 0.7589, 'grad_norm': 0.39739176630973816, 'learning_rate': 7.043035024376868e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.45, 'epoch': 0.55}

 63%|██████████████████████▌             | 771/1230 [5:50:22<3:27:21, 27.11s/it]
 63%|██████████████████████▌             | 772/1230 [5:50:47<3:22:27, 26.52s/it]
                                                                                
{'loss': 0.7071, 'grad_norm': 0.41894203424453735, 'learning_rate': 7.023102323669231e-05, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.34, 'epoch': 0.55}

 63%|██████████████████████▌             | 772/1230 [5:50:47<3:22:27, 26.52s/it]
 63%|██████████████████████▌             | 773/1230 [5:51:16<3:28:09, 27.33s/it]
                                                                                
{'loss': 0.7116, 'grad_norm': 0.397048681974411, 'learning_rate': 7.003189539564141e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.12, 'epoch': 0.55}

 63%|██████████████████████▌             | 773/1230 [5:51:16<3:28:09, 27.33s/it]
 63%|██████████████████████▋             | 774/1230 [5:51:42<3:23:18, 26.75s/it]
                                                                                
{'loss': 0.7018, 'grad_norm': 0.39036619663238525, 'learning_rate': 6.98329680991669e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.09, 'epoch': 0.55}

 63%|██████████████████████▋             | 774/1230 [5:51:42<3:23:18, 26.75s/it]
 63%|██████████████████████▋             | 775/1230 [5:52:07<3:19:44, 26.34s/it]
                                                                                
{'loss': 0.7025, 'grad_norm': 0.40661555528640747, 'learning_rate': 6.963424272443143e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.65, 'epoch': 0.55}

 63%|██████████████████████▋             | 775/1230 [5:52:07<3:19:44, 26.34s/it]
 63%|██████████████████████▋             | 776/1230 [5:52:37<3:26:46, 27.33s/it]
                                                                                
{'loss': 0.7321, 'grad_norm': 0.39608830213546753, 'learning_rate': 6.943572064719969e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.99, 'epoch': 0.55}

 63%|██████████████████████▋             | 776/1230 [5:52:37<3:26:46, 27.33s/it]
 63%|██████████████████████▋             | 777/1230 [5:53:02<3:22:47, 26.86s/it]
                                                                                
{'loss': 0.7672, 'grad_norm': 0.42595788836479187, 'learning_rate': 6.923740324182902e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.81, 'epoch': 0.55}

 63%|██████████████████████▋             | 777/1230 [5:53:02<3:22:47, 26.86s/it]
 63%|██████████████████████▊             | 778/1230 [5:53:28<3:18:42, 26.38s/it]
                                                                                
{'loss': 0.7211, 'grad_norm': 0.41514453291893005, 'learning_rate': 6.903929188125974e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.14, 'epoch': 0.55}

 63%|██████████████████████▊             | 778/1230 [5:53:28<3:18:42, 26.38s/it]
 63%|██████████████████████▊             | 779/1230 [5:53:54<3:18:00, 26.34s/it]
                                                                                
{'loss': 0.7438, 'grad_norm': 0.422076940536499, 'learning_rate': 6.884138793700576e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 175.99, 'epoch': 0.55}

 63%|██████████████████████▊             | 779/1230 [5:53:54<3:18:00, 26.34s/it]
 63%|██████████████████████▊             | 780/1230 [5:54:23<3:24:41, 27.29s/it]
                                                                                
{'loss': 0.7481, 'grad_norm': 0.40329769253730774, 'learning_rate': 6.864369277914509e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.22, 'epoch': 0.55}

 63%|██████████████████████▊             | 780/1230 [5:54:23<3:24:41, 27.29s/it]
 63%|██████████████████████▊             | 781/1230 [5:54:49<3:21:29, 26.93s/it]
                                                                                
{'loss': 0.7358, 'grad_norm': 0.39142659306526184, 'learning_rate': 6.844620777631035e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.33, 'epoch': 0.55}

 63%|██████████████████████▊             | 781/1230 [5:54:49<3:21:29, 26.93s/it]
 64%|██████████████████████▉             | 782/1230 [5:55:16<3:20:13, 26.82s/it]
                                                                                
{'loss': 0.7079, 'grad_norm': 0.3961651623249054, 'learning_rate': 6.824893429567921e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.07, 'epoch': 0.55}

 64%|██████████████████████▉             | 782/1230 [5:55:16<3:20:13, 26.82s/it]
 64%|██████████████████████▉             | 783/1230 [5:55:42<3:18:34, 26.65s/it]
                                                                                
{'loss': 0.7011, 'grad_norm': 0.3743172287940979, 'learning_rate': 6.805187370296496e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.41, 'epoch': 0.55}

 64%|██████████████████████▉             | 783/1230 [5:55:42<3:18:34, 26.65s/it]
 64%|██████████████████████▉             | 784/1230 [5:56:07<3:14:19, 26.14s/it]
                                                                                
{'loss': 0.733, 'grad_norm': 0.43782201409339905, 'learning_rate': 6.785502736240715e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.87, 'epoch': 0.55}

 64%|██████████████████████▉             | 784/1230 [5:56:07<3:14:19, 26.14s/it]
 64%|██████████████████████▉             | 785/1230 [5:56:36<3:20:13, 27.00s/it]
                                                                                
{'loss': 0.7793, 'grad_norm': 0.3936965763568878, 'learning_rate': 6.765839663676203e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.0, 'epoch': 0.56}

 64%|██████████████████████▉             | 785/1230 [5:56:36<3:20:13, 27.00s/it]
 64%|███████████████████████             | 786/1230 [5:57:02<3:17:35, 26.70s/it]
                                                                                
{'loss': 0.7406, 'grad_norm': 0.410284161567688, 'learning_rate': 6.746198288729319e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.64, 'epoch': 0.56}

 64%|███████████████████████             | 786/1230 [5:57:02<3:17:35, 26.70s/it]
 64%|███████████████████████             | 787/1230 [5:57:31<3:22:02, 27.36s/it]
                                                                                
{'loss': 0.747, 'grad_norm': 0.40015581250190735, 'learning_rate': 6.726578747376206e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.67, 'epoch': 0.56}

 64%|███████████████████████             | 787/1230 [5:57:31<3:22:02, 27.36s/it]
 64%|███████████████████████             | 788/1230 [5:58:01<3:26:09, 27.99s/it]
                                                                                
{'loss': 0.7165, 'grad_norm': 0.3892819881439209, 'learning_rate': 6.706981175441864e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.48, 'epoch': 0.56}

 64%|███████████████████████             | 788/1230 [5:58:01<3:26:09, 27.99s/it]
 64%|███████████████████████             | 789/1230 [5:58:26<3:19:59, 27.21s/it]
                                                                                
{'loss': 0.6992, 'grad_norm': 0.3884258568286896, 'learning_rate': 6.687405708599183e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.44, 'epoch': 0.56}

 64%|███████████████████████             | 789/1230 [5:58:26<3:19:59, 27.21s/it]
 64%|███████████████████████             | 790/1230 [5:58:51<3:15:34, 26.67s/it]
                                                                                
{'loss': 0.7664, 'grad_norm': 0.4248301088809967, 'learning_rate': 6.667852482368039e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.44, 'epoch': 0.56}

 64%|███████████████████████             | 790/1230 [5:58:51<3:15:34, 26.67s/it]
 64%|███████████████████████▏            | 791/1230 [5:59:21<3:21:35, 27.55s/it]
                                                                                
{'loss': 0.681, 'grad_norm': 0.40561971068382263, 'learning_rate': 6.648321632114326e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 147.65, 'epoch': 0.56}

 64%|███████████████████████▏            | 791/1230 [5:59:21<3:21:35, 27.55s/it]
 64%|███████████████████████▏            | 792/1230 [5:59:47<3:17:20, 27.03s/it]
                                                                                
{'loss': 0.7128, 'grad_norm': 0.4270016849040985, 'learning_rate': 6.628813293049033e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.25, 'epoch': 0.56}

 64%|███████████████████████▏            | 792/1230 [5:59:47<3:17:20, 27.03s/it]
 64%|███████████████████████▏            | 793/1230 [6:00:13<3:14:20, 26.68s/it]
                                                                                
{'loss': 0.7068, 'grad_norm': 0.38478633761405945, 'learning_rate': 6.609327600227311e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.68, 'epoch': 0.56}

 64%|███████████████████████▏            | 793/1230 [6:00:13<3:14:20, 26.68s/it]
 65%|███████████████████████▏            | 794/1230 [6:00:42<3:19:38, 27.47s/it]
                                                                                
{'loss': 0.7151, 'grad_norm': 0.3970317244529724, 'learning_rate': 6.589864688547521e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.44, 'epoch': 0.56}

 65%|███████████████████████▏            | 794/1230 [6:00:42<3:19:38, 27.47s/it]
 65%|███████████████████████▎            | 795/1230 [6:01:07<3:14:21, 26.81s/it]
                                                                                
{'loss': 0.7159, 'grad_norm': 0.4118540287017822, 'learning_rate': 6.570424692750318e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.53, 'epoch': 0.56}

 65%|███████████████████████▎            | 795/1230 [6:01:07<3:14:21, 26.81s/it]
 65%|███████████████████████▎            | 796/1230 [6:01:36<3:18:09, 27.39s/it]
                                                                                
{'loss': 0.7049, 'grad_norm': 0.3818812668323517, 'learning_rate': 6.551007747417715e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.97, 'epoch': 0.56}

 65%|███████████████████████▎            | 796/1230 [6:01:36<3:18:09, 27.39s/it]
 65%|███████████████████████▎            | 797/1230 [6:02:02<3:15:15, 27.06s/it]
                                                                                
{'loss': 0.7116, 'grad_norm': 0.39212653040885925, 'learning_rate': 6.53161398697214e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.37, 'epoch': 0.56}

 65%|███████████████████████▎            | 797/1230 [6:02:02<3:15:15, 27.06s/it]
 65%|███████████████████████▎            | 798/1230 [6:02:29<3:13:19, 26.85s/it]
                                                                                
{'loss': 0.7199, 'grad_norm': 0.402144193649292, 'learning_rate': 6.512243545675527e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.41, 'epoch': 0.56}

 65%|███████████████████████▎            | 798/1230 [6:02:29<3:13:19, 26.85s/it]
 65%|███████████████████████▍            | 799/1230 [6:02:54<3:09:04, 26.32s/it]
                                                                                
{'loss': 0.7336, 'grad_norm': 0.4236696660518646, 'learning_rate': 6.49289655762835e-05, 'memory/max_active (GiB)': 11.04, 'memory/max_allocated (GiB)': 11.04, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.01, 'epoch': 0.57}

 65%|███████████████████████▍            | 799/1230 [6:02:54<3:09:04, 26.32s/it]
 65%|███████████████████████▍            | 800/1230 [6:03:23<3:15:29, 27.28s/it]
                                                                                
{'loss': 0.7148, 'grad_norm': 0.3947851359844208, 'learning_rate': 6.473573156768737e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.43, 'epoch': 0.57}

 65%|███████████████████████▍            | 800/1230 [6:03:23<3:15:29, 27.28s/it]
 65%|███████████████████████▍            | 801/1230 [6:03:49<3:12:53, 26.98s/it]
                                                                                
{'loss': 0.7472, 'grad_norm': 0.45719990134239197, 'learning_rate': 6.454273476871518e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.6, 'epoch': 0.57}

 65%|███████████████████████▍            | 801/1230 [6:03:50<3:12:53, 26.98s/it]
 65%|███████████████████████▍            | 802/1230 [6:04:14<3:08:01, 26.36s/it]
                                                                                
{'loss': 0.7038, 'grad_norm': 0.40754398703575134, 'learning_rate': 6.434997651547303e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.31, 'epoch': 0.57}

 65%|███████████████████████▍            | 802/1230 [6:04:14<3:08:01, 26.36s/it]
 65%|███████████████████████▌            | 803/1230 [6:04:39<3:04:35, 25.94s/it]
                                                                                
{'loss': 0.6921, 'grad_norm': 0.3760529160499573, 'learning_rate': 6.41574581424156e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.88, 'epoch': 0.57}

 65%|███████████████████████▌            | 803/1230 [6:04:39<3:04:35, 25.94s/it]
 65%|███████████████████████▌            | 804/1230 [6:05:04<3:01:51, 25.61s/it]
                                                                                
{'loss': 0.6992, 'grad_norm': 0.422107994556427, 'learning_rate': 6.39651809823368e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.43, 'epoch': 0.57}

 65%|███████████████████████▌            | 804/1230 [6:05:04<3:01:51, 25.61s/it]
 65%|███████████████████████▌            | 805/1230 [6:05:30<3:02:11, 25.72s/it]
                                                                                
{'loss': 0.7483, 'grad_norm': 0.44880911707878113, 'learning_rate': 6.377314636636078e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.38, 'epoch': 0.57}

 65%|███████████████████████▌            | 805/1230 [6:05:30<3:02:11, 25.72s/it]
 66%|███████████████████████▌            | 806/1230 [6:05:55<3:00:01, 25.47s/it]
                                                                                
{'loss': 0.7236, 'grad_norm': 0.40190431475639343, 'learning_rate': 6.358135562393247e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.09, 'epoch': 0.57}

 66%|███████████████████████▌            | 806/1230 [6:05:55<3:00:01, 25.47s/it]
 66%|███████████████████████▌            | 807/1230 [6:06:20<2:58:39, 25.34s/it]
                                                                                
{'loss': 0.7332, 'grad_norm': 0.4240853786468506, 'learning_rate': 6.338981008280854e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.57, 'epoch': 0.57}

 66%|███████████████████████▌            | 807/1230 [6:06:20<2:58:39, 25.34s/it]
 66%|███████████████████████▋            | 808/1230 [6:06:46<3:00:10, 25.62s/it]
                                                                                
{'loss': 0.7856, 'grad_norm': 0.4556266665458679, 'learning_rate': 6.31985110690481e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.78, 'epoch': 0.57}

 66%|███████████████████████▋            | 808/1230 [6:06:46<3:00:10, 25.62s/it]
 66%|███████████████████████▋            | 809/1230 [6:07:15<3:07:00, 26.65s/it]
                                                                                
{'loss': 0.69, 'grad_norm': 0.41872039437294006, 'learning_rate': 6.300745990700357e-05, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.47, 'epoch': 0.57}

 66%|███████████████████████▋            | 809/1230 [6:07:15<3:07:00, 26.65s/it]
 66%|███████████████████████▋            | 810/1230 [6:07:44<3:11:24, 27.34s/it]
                                                                                
{'loss': 0.7042, 'grad_norm': 0.38044771552085876, 'learning_rate': 6.281665791931151e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.54, 'epoch': 0.57}

 66%|███████████████████████▋            | 810/1230 [6:07:44<3:11:24, 27.34s/it]
 66%|███████████████████████▋            | 811/1230 [6:08:10<3:08:15, 26.96s/it]
                                                                                
{'loss': 0.7152, 'grad_norm': 0.4044669568538666, 'learning_rate': 6.262610642688348e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.47, 'epoch': 0.57}

 66%|███████████████████████▋            | 811/1230 [6:08:10<3:08:15, 26.96s/it]
 66%|███████████████████████▊            | 812/1230 [6:08:36<3:04:32, 26.49s/it]
                                                                                
{'loss': 0.6934, 'grad_norm': 0.44398418068885803, 'learning_rate': 6.243580674889683e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.5, 'epoch': 0.57}

 66%|███████████████████████▊            | 812/1230 [6:08:36<3:04:32, 26.49s/it]
 66%|███████████████████████▊            | 813/1230 [6:09:01<3:01:47, 26.16s/it]
                                                                                
{'loss': 0.7187, 'grad_norm': 0.40107762813568115, 'learning_rate': 6.224576020278572e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.07, 'epoch': 0.58}

 66%|███████████████████████▊            | 813/1230 [6:09:01<3:01:47, 26.16s/it]
 66%|███████████████████████▊            | 814/1230 [6:09:31<3:08:43, 27.22s/it]
                                                                                
{'loss': 0.7794, 'grad_norm': 0.4375170171260834, 'learning_rate': 6.205596810423171e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.81, 'epoch': 0.58}

 66%|███████████████████████▊            | 814/1230 [6:09:31<3:08:43, 27.22s/it]
 66%|███████████████████████▊            | 815/1230 [6:09:56<3:03:49, 26.58s/it]
                                                                                
{'loss': 0.7243, 'grad_norm': 0.41030392050743103, 'learning_rate': 6.186643176715501e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.69, 'epoch': 0.58}

 66%|███████████████████████▊            | 815/1230 [6:09:56<3:03:49, 26.58s/it]
 66%|███████████████████████▉            | 816/1230 [6:10:21<3:00:05, 26.10s/it]
                                                                                
{'loss': 0.668, 'grad_norm': 0.4074665307998657, 'learning_rate': 6.16771525037051e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.43, 'epoch': 0.58}

 66%|███████████████████████▉            | 816/1230 [6:10:21<3:00:05, 26.10s/it]
 66%|███████████████████████▉            | 817/1230 [6:10:47<2:58:58, 26.00s/it]
                                                                                
{'loss': 0.7667, 'grad_norm': 0.43760228157043457, 'learning_rate': 6.148813162425184e-05, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.44, 'epoch': 0.58}

 66%|███████████████████████▉            | 817/1230 [6:10:47<2:58:58, 26.00s/it]
 67%|███████████████████████▉            | 818/1230 [6:11:13<2:59:25, 26.13s/it]
                                                                                
{'loss': 0.7096, 'grad_norm': 0.391703337430954, 'learning_rate': 6.12993704373762e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.92, 'epoch': 0.58}

 67%|███████████████████████▉            | 818/1230 [6:11:13<2:59:25, 26.13s/it]
 67%|███████████████████████▉            | 819/1230 [6:11:42<3:05:22, 27.06s/it]
                                                                                
{'loss': 0.7203, 'grad_norm': 0.38580092787742615, 'learning_rate': 6.111087024986139e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.62, 'epoch': 0.58}

 67%|███████████████████████▉            | 819/1230 [6:11:42<3:05:22, 27.06s/it]
 67%|████████████████████████            | 820/1230 [6:12:11<3:08:24, 27.57s/it]
                                                                                
{'loss': 0.7006, 'grad_norm': 0.40392985939979553, 'learning_rate': 6.0922632366683774e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.5, 'epoch': 0.58}

 67%|████████████████████████            | 820/1230 [6:12:11<3:08:24, 27.57s/it]
 67%|████████████████████████            | 821/1230 [6:12:40<3:10:43, 27.98s/it]
                                                                                
{'loss': 0.6921, 'grad_norm': 0.4298458397388458, 'learning_rate': 6.073465809100367e-05, 'memory/max_active (GiB)': 10.52, 'memory/max_allocated (GiB)': 10.52, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 147.24, 'epoch': 0.58}

 67%|████████████████████████            | 821/1230 [6:12:40<3:10:43, 27.98s/it]
 67%|████████████████████████            | 822/1230 [6:13:09<3:12:49, 28.36s/it]
                                                                                
{'loss': 0.7205, 'grad_norm': 0.4135745167732239, 'learning_rate': 6.054694872415657e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.2, 'epoch': 0.58}

 67%|████████████████████████            | 822/1230 [6:13:09<3:12:49, 28.36s/it]
 67%|████████████████████████            | 823/1230 [6:13:35<3:05:50, 27.40s/it]
                                                                                
{'loss': 0.7609, 'grad_norm': 0.42149507999420166, 'learning_rate': 6.035950556564397e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.29, 'epoch': 0.58}

 67%|████████████████████████            | 823/1230 [6:13:35<3:05:50, 27.40s/it]
 67%|████████████████████████            | 824/1230 [6:14:00<3:01:30, 26.82s/it]
                                                                                
{'loss': 0.6954, 'grad_norm': 0.3767984211444855, 'learning_rate': 6.017232991312442e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.83, 'epoch': 0.58}

 67%|████████████████████████            | 824/1230 [6:14:00<3:01:30, 26.82s/it]
 67%|████████████████████████▏           | 825/1230 [6:14:26<2:58:45, 26.48s/it]
                                                                                
{'loss': 0.7122, 'grad_norm': 0.39049819111824036, 'learning_rate': 5.998542306240454e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.63, 'epoch': 0.58}

 67%|████████████████████████▏           | 825/1230 [6:14:26<2:58:45, 26.48s/it]
 67%|████████████████████████▏           | 826/1230 [6:14:55<3:04:33, 27.41s/it]
                                                                                
{'loss': 0.7258, 'grad_norm': 0.39489367604255676, 'learning_rate': 5.9798786307430075e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.62, 'epoch': 0.58}

 67%|████████████████████████▏           | 826/1230 [6:14:55<3:04:33, 27.41s/it]
 67%|████████████████████████▏           | 827/1230 [6:15:25<3:08:47, 28.11s/it]
                                                                                
{'loss': 0.7091, 'grad_norm': 0.4039662778377533, 'learning_rate': 5.961242094027688e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.18, 'epoch': 0.59}

 67%|████████████████████████▏           | 827/1230 [6:15:25<3:08:47, 28.11s/it]
 67%|████████████████████████▏           | 828/1230 [6:15:55<3:11:22, 28.56s/it]
                                                                                
{'loss': 0.7233, 'grad_norm': 0.39788731932640076, 'learning_rate': 5.942632825114208e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.47, 'epoch': 0.59}

 67%|████████████████████████▏           | 828/1230 [6:15:55<3:11:22, 28.56s/it]
 67%|████████████████████████▎           | 829/1230 [6:16:24<3:12:44, 28.84s/it]
                                                                                
{'loss': 0.726, 'grad_norm': 0.39901652932167053, 'learning_rate': 5.924050952833493e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.57, 'epoch': 0.59}

 67%|████████████████████████▎           | 829/1230 [6:16:24<3:12:44, 28.84s/it]
 67%|████████████████████████▎           | 830/1230 [6:16:53<3:13:04, 28.96s/it]
                                                                                
{'loss': 0.7079, 'grad_norm': 0.3764503002166748, 'learning_rate': 5.905496605826813e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.9, 'epoch': 0.59}

 67%|████████████████████████▎           | 830/1230 [6:16:53<3:13:04, 28.96s/it]
 68%|████████████████████████▎           | 831/1230 [6:17:23<3:13:09, 29.05s/it]
                                                                                
{'loss': 0.7338, 'grad_norm': 0.4030141830444336, 'learning_rate': 5.8869699125448826e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.39, 'epoch': 0.59}

 68%|████████████████████████▎           | 831/1230 [6:17:23<3:13:09, 29.05s/it]
 68%|████████████████████████▎           | 832/1230 [6:17:52<3:13:21, 29.15s/it]
                                                                                
{'loss': 0.6809, 'grad_norm': 0.39091554284095764, 'learning_rate': 5.868471001246967e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.83, 'epoch': 0.59}

 68%|████████████████████████▎           | 832/1230 [6:17:52<3:13:21, 29.15s/it]
 68%|████████████████████████▍           | 833/1230 [6:18:17<3:05:14, 28.00s/it]
                                                                                
{'loss': 0.6981, 'grad_norm': 0.4093121290206909, 'learning_rate': 5.8500000000000026e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.78, 'epoch': 0.59}

 68%|████████████████████████▍           | 833/1230 [6:18:17<3:05:14, 28.00s/it]
 68%|████████████████████████▍           | 834/1230 [6:18:47<3:08:10, 28.51s/it]
                                                                                
{'loss': 0.6948, 'grad_norm': 0.39781418442726135, 'learning_rate': 5.8315570366776975e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.18, 'epoch': 0.59}

 68%|████████████████████████▍           | 834/1230 [6:18:47<3:08:10, 28.51s/it]
 68%|████████████████████████▍           | 835/1230 [6:19:16<3:09:23, 28.77s/it]
                                                                                
{'loss': 0.748, 'grad_norm': 0.4110676348209381, 'learning_rate': 5.813142238959664e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.21, 'epoch': 0.59}

 68%|████████████████████████▍           | 835/1230 [6:19:16<3:09:23, 28.77s/it]
 68%|████████████████████████▍           | 836/1230 [6:19:45<3:09:11, 28.81s/it]
                                                                                
{'loss': 0.7029, 'grad_norm': 0.3987717032432556, 'learning_rate': 5.7947557343305215e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.91, 'epoch': 0.59}

 68%|████████████████████████▍           | 836/1230 [6:19:45<3:09:11, 28.81s/it]
 68%|████████████████████████▍           | 837/1230 [6:20:11<3:02:01, 27.79s/it]
                                                                                
{'loss': 0.7232, 'grad_norm': 0.41662612557411194, 'learning_rate': 5.776397650079019e-05, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 203.06, 'epoch': 0.59}

 68%|████████████████████████▍           | 837/1230 [6:20:11<3:02:01, 27.79s/it]
 68%|████████████████████████▌           | 838/1230 [6:20:40<3:03:33, 28.10s/it]
                                                                                
{'loss': 0.6991, 'grad_norm': 0.37876400351524353, 'learning_rate': 5.758068113297156e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.05, 'epoch': 0.59}

 68%|████████████████████████▌           | 838/1230 [6:20:40<3:03:33, 28.10s/it]
 68%|████████████████████████▌           | 839/1230 [6:21:09<3:05:16, 28.43s/it]
                                                                                
{'loss': 0.6768, 'grad_norm': 0.36448565125465393, 'learning_rate': 5.7397672508792875e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.36, 'epoch': 0.59}

 68%|████████████████████████▌           | 839/1230 [6:21:09<3:05:16, 28.43s/it]
 68%|████████████████████████▌           | 840/1230 [6:21:35<3:00:22, 27.75s/it]
                                                                                
{'loss': 0.7302, 'grad_norm': 0.42157360911369324, 'learning_rate': 5.7214951895212716e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.02, 'epoch': 0.59}

 68%|████████████████████████▌           | 840/1230 [6:21:35<3:00:22, 27.75s/it]
 68%|████████████████████████▌           | 841/1230 [6:22:04<3:02:32, 28.16s/it]
                                                                                
{'loss': 0.7295, 'grad_norm': 0.41280245780944824, 'learning_rate': 5.703252055719572e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.01, 'epoch': 0.6}

 68%|████████████████████████▌           | 841/1230 [6:22:04<3:02:32, 28.16s/it]
 68%|████████████████████████▋           | 842/1230 [6:22:29<2:56:48, 27.34s/it]
                                                                                
{'loss': 0.687, 'grad_norm': 0.3938748836517334, 'learning_rate': 5.685037975770388e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.04, 'epoch': 0.6}

 68%|████████████████████████▋           | 842/1230 [6:22:29<2:56:48, 27.34s/it]
 69%|████████████████████████▋           | 843/1230 [6:22:59<3:00:55, 28.05s/it]
                                                                                
{'loss': 0.6634, 'grad_norm': 0.3966750204563141, 'learning_rate': 5.666853075768784e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.37, 'epoch': 0.6}

 69%|████████████████████████▋           | 843/1230 [6:22:59<3:00:55, 28.05s/it]
 69%|████████████████████████▋           | 844/1230 [6:23:25<2:55:27, 27.27s/it]
                                                                                
{'loss': 0.6869, 'grad_norm': 0.3965831398963928, 'learning_rate': 5.648697481607807e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.77, 'epoch': 0.6}

 69%|████████████████████████▋           | 844/1230 [6:23:25<2:55:27, 27.27s/it]
 69%|████████████████████████▋           | 845/1230 [6:23:50<2:51:04, 26.66s/it]
                                                                                
{'loss': 0.6967, 'grad_norm': 0.4248439073562622, 'learning_rate': 5.630571318977626e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.57, 'epoch': 0.6}

 69%|████████████████████████▋           | 845/1230 [6:23:50<2:51:04, 26.66s/it]
 69%|████████████████████████▊           | 846/1230 [6:24:19<2:55:54, 27.49s/it]
                                                                                
{'loss': 0.6733, 'grad_norm': 0.38163578510284424, 'learning_rate': 5.612474713364662e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.71, 'epoch': 0.6}

 69%|████████████████████████▊           | 846/1230 [6:24:19<2:55:54, 27.49s/it]
 69%|████████████████████████▊           | 847/1230 [6:24:46<2:53:08, 27.12s/it]
                                                                                
{'loss': 0.7028, 'grad_norm': 0.41237446665763855, 'learning_rate': 5.5944077900507e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.29, 'epoch': 0.6}

 69%|████████████████████████▊           | 847/1230 [6:24:46<2:53:08, 27.12s/it]
 69%|████████████████████████▊           | 848/1230 [6:25:12<2:51:48, 26.99s/it]
                                                                                
{'loss': 0.7148, 'grad_norm': 0.42833253741264343, 'learning_rate': 5.576370674112049e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.84, 'epoch': 0.6}

 69%|████████████████████████▊           | 848/1230 [6:25:12<2:51:48, 26.99s/it]
 69%|████████████████████████▊           | 849/1230 [6:25:38<2:48:09, 26.48s/it]
                                                                                
{'loss': 0.7148, 'grad_norm': 0.42638927698135376, 'learning_rate': 5.558363490418662e-05, 'memory/max_active (GiB)': 9.76, 'memory/max_allocated (GiB)': 9.76, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.65, 'epoch': 0.6}

 69%|████████████████████████▊           | 849/1230 [6:25:38<2:48:09, 26.48s/it]
 69%|████████████████████████▉           | 850/1230 [6:26:02<2:44:36, 25.99s/it]
                                                                                
{'loss': 0.6629, 'grad_norm': 0.3738210201263428, 'learning_rate': 5.5403863636332646e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.74, 'epoch': 0.6}

 69%|████████████████████████▉           | 850/1230 [6:26:02<2:44:36, 25.99s/it]
 69%|████████████████████████▉           | 851/1230 [6:26:31<2:49:31, 26.84s/it]
                                                                                
{'loss': 0.7343, 'grad_norm': 0.3898461163043976, 'learning_rate': 5.522439418210507e-05, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.52, 'epoch': 0.6}

 69%|████████████████████████▉           | 851/1230 [6:26:31<2:49:31, 26.84s/it]
 69%|████████████████████████▉           | 852/1230 [6:26:57<2:47:31, 26.59s/it]
                                                                                
{'loss': 0.7124, 'grad_norm': 0.42209020256996155, 'learning_rate': 5.5045227783960956e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.42, 'epoch': 0.6}

 69%|████████████████████████▉           | 852/1230 [6:26:57<2:47:31, 26.59s/it]
 69%|████████████████████████▉           | 853/1230 [6:27:26<2:51:17, 27.26s/it]
                                                                                
{'loss': 0.6945, 'grad_norm': 0.3819439113140106, 'learning_rate': 5.486636568225932e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.04, 'epoch': 0.6}

 69%|████████████████████████▉           | 853/1230 [6:27:26<2:51:17, 27.26s/it]
 69%|████████████████████████▉           | 854/1230 [6:27:51<2:47:09, 26.67s/it]
                                                                                
{'loss': 0.6884, 'grad_norm': 0.3897852897644043, 'learning_rate': 5.468780911525249e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.47, 'epoch': 0.6}

 69%|████████████████████████▉           | 854/1230 [6:27:51<2:47:09, 26.67s/it]
 70%|█████████████████████████           | 855/1230 [6:28:21<2:51:50, 27.49s/it]
                                                                                
{'loss': 0.6828, 'grad_norm': 0.37862205505371094, 'learning_rate': 5.450955931907768e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.99, 'epoch': 0.61}

 70%|█████████████████████████           | 855/1230 [6:28:21<2:51:50, 27.49s/it]
 70%|█████████████████████████           | 856/1230 [6:28:50<2:55:22, 28.13s/it]
                                                                                
{'loss': 0.7078, 'grad_norm': 0.41848045587539673, 'learning_rate': 5.4331617527748296e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.28, 'epoch': 0.61}

 70%|█████████████████████████           | 856/1230 [6:28:50<2:55:22, 28.13s/it]
 70%|█████████████████████████           | 857/1230 [6:29:17<2:51:36, 27.61s/it]
                                                                                
{'loss': 0.7037, 'grad_norm': 0.41892585158348083, 'learning_rate': 5.415398497314548e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.23, 'epoch': 0.61}

 70%|█████████████████████████           | 857/1230 [6:29:17<2:51:36, 27.61s/it]
 70%|█████████████████████████           | 858/1230 [6:29:46<2:54:25, 28.13s/it]
                                                                                
{'loss': 0.6908, 'grad_norm': 0.40383094549179077, 'learning_rate': 5.397666288500954e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.05, 'epoch': 0.61}

 70%|█████████████████████████           | 858/1230 [6:29:46<2:54:25, 28.13s/it]
 70%|█████████████████████████▏          | 859/1230 [6:30:11<2:48:04, 27.18s/it]
                                                                                
{'loss': 0.7042, 'grad_norm': 0.39924341440200806, 'learning_rate': 5.379965249093136e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.66, 'epoch': 0.61}

 70%|█████████████████████████▏          | 859/1230 [6:30:11<2:48:04, 27.18s/it]
 70%|█████████████████████████▏          | 860/1230 [6:30:36<2:43:45, 26.55s/it]
                                                                                
{'loss': 0.7356, 'grad_norm': 0.4356042742729187, 'learning_rate': 5.36229550163441e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.8, 'epoch': 0.61}

 70%|█████████████████████████▏          | 860/1230 [6:30:36<2:43:45, 26.55s/it]
 70%|█████████████████████████▏          | 861/1230 [6:31:02<2:42:30, 26.43s/it]
                                                                                
{'loss': 0.7495, 'grad_norm': 0.41518282890319824, 'learning_rate': 5.344657168451453e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.02, 'epoch': 0.61}

 70%|█████████████████████████▏          | 861/1230 [6:31:02<2:42:30, 26.43s/it]
 70%|█████████████████████████▏          | 862/1230 [6:31:29<2:41:44, 26.37s/it]
                                                                                
{'loss': 0.704, 'grad_norm': 0.4009655714035034, 'learning_rate': 5.3270503716534626e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.74, 'epoch': 0.61}

 70%|█████████████████████████▏          | 862/1230 [6:31:29<2:41:44, 26.37s/it]
 70%|█████████████████████████▎          | 863/1230 [6:31:57<2:45:55, 27.13s/it]
                                                                                
{'loss': 0.7366, 'grad_norm': 0.4159681499004364, 'learning_rate': 5.30947523313132e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.34, 'epoch': 0.61}

 70%|█████████████████████████▎          | 863/1230 [6:31:57<2:45:55, 27.13s/it]
 70%|█████████████████████████▎          | 864/1230 [6:32:26<2:48:42, 27.66s/it]
                                                                                
{'loss': 0.708, 'grad_norm': 0.38671955466270447, 'learning_rate': 5.2919318745567246e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.1, 'epoch': 0.61}

 70%|█████████████████████████▎          | 864/1230 [6:32:26<2:48:42, 27.66s/it]
 70%|█████████████████████████▎          | 865/1230 [6:32:55<2:51:00, 28.11s/it]
                                                                                
{'loss': 0.7346, 'grad_norm': 0.4159725606441498, 'learning_rate': 5.2744204173813736e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.04, 'epoch': 0.61}

 70%|█████████████████████████▎          | 865/1230 [6:32:55<2:51:00, 28.11s/it]
 70%|█████████████████████████▎          | 866/1230 [6:33:25<2:52:54, 28.50s/it]
                                                                                
{'loss': 0.6822, 'grad_norm': 0.39667025208473206, 'learning_rate': 5.25694098283611e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.57, 'epoch': 0.61}

 70%|█████████████████████████▎          | 866/1230 [6:33:25<2:52:54, 28.50s/it]
 70%|█████████████████████████▍          | 867/1230 [6:33:50<2:46:33, 27.53s/it]
                                                                                
{'loss': 0.7114, 'grad_norm': 0.4290774464607239, 'learning_rate': 5.2394936919300884e-05, 'memory/max_active (GiB)': 10.52, 'memory/max_allocated (GiB)': 10.52, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.02, 'epoch': 0.61}

 70%|█████████████████████████▍          | 867/1230 [6:33:50<2:46:33, 27.53s/it]
 71%|█████████████████████████▍          | 868/1230 [6:34:16<2:42:45, 26.98s/it]
                                                                                
{'loss': 0.7039, 'grad_norm': 0.40482595562934875, 'learning_rate': 5.2220786654499344e-05, 'memory/max_active (GiB)': 7.96, 'memory/max_allocated (GiB)': 7.96, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 211.89, 'epoch': 0.61}

 71%|█████████████████████████▍          | 868/1230 [6:34:16<2:42:45, 26.98s/it]
 71%|█████████████████████████▍          | 869/1230 [6:34:45<2:46:54, 27.74s/it]
                                                                                
{'loss': 0.7011, 'grad_norm': 0.39133018255233765, 'learning_rate': 5.2046960239589024e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.71, 'epoch': 0.62}

 71%|█████████████████████████▍          | 869/1230 [6:34:45<2:46:54, 27.74s/it]
 71%|█████████████████████████▍          | 870/1230 [6:35:10<2:41:41, 26.95s/it]
                                                                                
{'loss': 0.6718, 'grad_norm': 0.41495901346206665, 'learning_rate': 5.187345887796055e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.35, 'epoch': 0.62}

 71%|█████████████████████████▍          | 870/1230 [6:35:10<2:41:41, 26.95s/it]
 71%|█████████████████████████▍          | 871/1230 [6:35:40<2:45:42, 27.69s/it]
                                                                                
{'loss': 0.6974, 'grad_norm': 0.4004170000553131, 'learning_rate': 5.1700283770754185e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.37, 'epoch': 0.62}

 71%|█████████████████████████▍          | 871/1230 [6:35:40<2:45:42, 27.69s/it]
 71%|█████████████████████████▌          | 872/1230 [6:36:05<2:41:21, 27.04s/it]
                                                                                
{'loss': 0.7037, 'grad_norm': 0.39344286918640137, 'learning_rate': 5.152743611685159e-05, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.78, 'epoch': 0.62}

 71%|█████████████████████████▌          | 872/1230 [6:36:05<2:41:21, 27.04s/it]
 71%|█████████████████████████▌          | 873/1230 [6:36:30<2:37:20, 26.44s/it]
                                                                                
{'loss': 0.6805, 'grad_norm': 0.37588921189308167, 'learning_rate': 5.135491711286745e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.3, 'epoch': 0.62}

 71%|█████████████████████████▌          | 873/1230 [6:36:30<2:37:20, 26.44s/it]
 71%|█████████████████████████▌          | 874/1230 [6:36:56<2:34:27, 26.03s/it]
                                                                                
{'loss': 0.7113, 'grad_norm': 0.39842236042022705, 'learning_rate': 5.11827279531412e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.4, 'epoch': 0.62}

 71%|█████████████████████████▌          | 874/1230 [6:36:56<2:34:27, 26.03s/it]
 71%|█████████████████████████▌          | 875/1230 [6:37:21<2:33:51, 26.00s/it]
                                                                                
{'loss': 0.7105, 'grad_norm': 0.405127614736557, 'learning_rate': 5.1010869829728835e-05, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.53, 'epoch': 0.62}

 71%|█████████████████████████▌          | 875/1230 [6:37:21<2:33:51, 26.00s/it]
 71%|█████████████████████████▋          | 876/1230 [6:37:46<2:31:34, 25.69s/it]
                                                                                
{'loss': 0.6948, 'grad_norm': 0.415200799703598, 'learning_rate': 5.0839343932394594e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.77, 'epoch': 0.62}

 71%|█████████████████████████▋          | 876/1230 [6:37:46<2:31:34, 25.69s/it]
 71%|█████████████████████████▋          | 877/1230 [6:38:12<2:31:41, 25.78s/it]
                                                                                
{'loss': 0.7021, 'grad_norm': 0.3951016068458557, 'learning_rate': 5.066815144860267e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.22, 'epoch': 0.62}

 71%|█████████████████████████▋          | 877/1230 [6:38:12<2:31:41, 25.78s/it]
 71%|█████████████████████████▋          | 878/1230 [6:38:41<2:36:44, 26.72s/it]
                                                                                
{'loss': 0.7295, 'grad_norm': 0.39302584528923035, 'learning_rate': 5.049729356350918e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.55, 'epoch': 0.62}

 71%|█████████████████████████▋          | 878/1230 [6:38:41<2:36:44, 26.72s/it]
 71%|█████████████████████████▋          | 879/1230 [6:39:06<2:33:29, 26.24s/it]
                                                                                
{'loss': 0.6819, 'grad_norm': 0.4316733479499817, 'learning_rate': 5.0326771459953684e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.5, 'epoch': 0.62}

 71%|█████████████████████████▋          | 879/1230 [6:39:06<2:33:29, 26.24s/it]
 72%|█████████████████████████▊          | 880/1230 [6:39:32<2:31:04, 25.90s/it]
                                                                                
{'loss': 0.7153, 'grad_norm': 0.4076690971851349, 'learning_rate': 5.015658631845124e-05, 'memory/max_active (GiB)': 9.38, 'memory/max_allocated (GiB)': 9.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.05, 'epoch': 0.62}

 72%|█████████████████████████▊          | 880/1230 [6:39:32<2:31:04, 25.90s/it]
 72%|█████████████████████████▊          | 881/1230 [6:39:57<2:29:06, 25.63s/it]
                                                                                
{'loss': 0.6968, 'grad_norm': 0.38317129015922546, 'learning_rate': 4.9986739317184125e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.82, 'epoch': 0.62}

 72%|█████████████████████████▊          | 881/1230 [6:39:57<2:29:06, 25.63s/it]
 72%|█████████████████████████▊          | 882/1230 [6:40:22<2:29:04, 25.70s/it]
                                                                                
{'loss': 0.6951, 'grad_norm': 0.4106966555118561, 'learning_rate': 4.9817231631993663e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 200.25, 'epoch': 0.62}

 72%|█████████████████████████▊          | 882/1230 [6:40:22<2:29:04, 25.70s/it]
 72%|█████████████████████████▊          | 883/1230 [6:40:47<2:27:21, 25.48s/it]
                                                                                
{'loss': 0.6914, 'grad_norm': 0.39762619137763977, 'learning_rate': 4.9648064436372175e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.86, 'epoch': 0.62}

 72%|█████████████████████████▊          | 883/1230 [6:40:47<2:27:21, 25.48s/it]
 72%|█████████████████████████▊          | 884/1230 [6:41:14<2:28:15, 25.71s/it]
                                                                                
{'loss': 0.698, 'grad_norm': 0.43110501766204834, 'learning_rate': 4.947923890145467e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.34, 'epoch': 0.63}

 72%|█████████████████████████▊          | 884/1230 [6:41:14<2:28:15, 25.71s/it]
 72%|█████████████████████████▉          | 885/1230 [6:41:43<2:34:04, 26.80s/it]
                                                                                
{'loss': 0.6611, 'grad_norm': 0.4004526138305664, 'learning_rate': 4.931075619601098e-05, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.09, 'epoch': 0.63}

 72%|█████████████████████████▉          | 885/1230 [6:41:43<2:34:04, 26.80s/it]
 72%|█████████████████████████▉          | 886/1230 [6:42:08<2:30:46, 26.30s/it]
                                                                                
{'loss': 0.7021, 'grad_norm': 0.41781365871429443, 'learning_rate': 4.91426174864375e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.73, 'epoch': 0.63}

 72%|█████████████████████████▉          | 886/1230 [6:42:08<2:30:46, 26.30s/it]
 72%|█████████████████████████▉          | 887/1230 [6:42:33<2:28:12, 25.93s/it]
                                                                                
{'loss': 0.7328, 'grad_norm': 0.4008306562900543, 'learning_rate': 4.897482393674914e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.79, 'epoch': 0.63}

 72%|█████████████████████████▉          | 887/1230 [6:42:33<2:28:12, 25.93s/it]
 72%|█████████████████████████▉          | 888/1230 [6:42:58<2:26:39, 25.73s/it]
                                                                                
{'loss': 0.6917, 'grad_norm': 0.4029233157634735, 'learning_rate': 4.880737670857134e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.79, 'epoch': 0.63}

 72%|█████████████████████████▉          | 888/1230 [6:42:58<2:26:39, 25.73s/it]
 72%|██████████████████████████          | 889/1230 [6:43:28<2:32:39, 26.86s/it]
                                                                                
{'loss': 0.6794, 'grad_norm': 0.4029139280319214, 'learning_rate': 4.864027696113191e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.51, 'epoch': 0.63}

 72%|██████████████████████████          | 889/1230 [6:43:28<2:32:39, 26.86s/it]
 72%|██████████████████████████          | 890/1230 [6:43:57<2:36:32, 27.63s/it]
                                                                                
{'loss': 0.687, 'grad_norm': 0.40900301933288574, 'learning_rate': 4.8473525851253114e-05, 'memory/max_active (GiB)': 11.04, 'memory/max_allocated (GiB)': 11.04, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.33, 'epoch': 0.63}

 72%|██████████████████████████          | 890/1230 [6:43:57<2:36:32, 27.63s/it]
 72%|██████████████████████████          | 891/1230 [6:44:24<2:33:43, 27.21s/it]
                                                                                
{'loss': 0.6878, 'grad_norm': 0.42954203486442566, 'learning_rate': 4.83071245333436e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.61, 'epoch': 0.63}

 72%|██████████████████████████          | 891/1230 [6:44:24<2:33:43, 27.21s/it]
 73%|██████████████████████████          | 892/1230 [6:44:53<2:36:17, 27.74s/it]
                                                                                
{'loss': 0.7036, 'grad_norm': 0.3904794752597809, 'learning_rate': 4.814107415939047e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.18, 'epoch': 0.63}

 73%|██████████████████████████          | 892/1230 [6:44:53<2:36:17, 27.74s/it]
 73%|██████████████████████████▏         | 893/1230 [6:45:22<2:37:57, 28.12s/it]
                                                                                
{'loss': 0.7221, 'grad_norm': 0.387525737285614, 'learning_rate': 4.7975375878951215e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.1, 'epoch': 0.63}

 73%|██████████████████████████▏         | 893/1230 [6:45:22<2:37:57, 28.12s/it]
 73%|██████████████████████████▏         | 894/1230 [6:45:51<2:39:20, 28.45s/it]
                                                                                
{'loss': 0.7368, 'grad_norm': 0.43186259269714355, 'learning_rate': 4.781003083914576e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.66, 'epoch': 0.63}

 73%|██████████████████████████▏         | 894/1230 [6:45:51<2:39:20, 28.45s/it]
 73%|██████████████████████████▏         | 895/1230 [6:46:16<2:32:50, 27.37s/it]
                                                                                
{'loss': 0.6674, 'grad_norm': 0.4098013937473297, 'learning_rate': 4.7645040184648674e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.73, 'epoch': 0.63}

 73%|██████████████████████████▏         | 895/1230 [6:46:16<2:32:50, 27.37s/it]
 73%|██████████████████████████▏         | 896/1230 [6:46:41<2:28:26, 26.67s/it]
                                                                                
{'loss': 0.6806, 'grad_norm': 0.40792331099510193, 'learning_rate': 4.748040505768104e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.9, 'epoch': 0.63}

 73%|██████████████████████████▏         | 896/1230 [6:46:41<2:28:26, 26.67s/it]
 73%|██████████████████████████▎         | 897/1230 [6:47:06<2:25:20, 26.19s/it]
                                                                                
{'loss': 0.7073, 'grad_norm': 0.4042823016643524, 'learning_rate': 4.731612659800271e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.27, 'epoch': 0.63}

 73%|██████████████████████████▎         | 897/1230 [6:47:06<2:25:20, 26.19s/it]
 73%|██████████████████████████▎         | 898/1230 [6:47:31<2:22:54, 25.83s/it]
                                                                                
{'loss': 0.7385, 'grad_norm': 0.3959033191204071, 'learning_rate': 4.715220594290432e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.91, 'epoch': 0.64}

 73%|██████████████████████████▎         | 898/1230 [6:47:31<2:22:54, 25.83s/it]
 73%|██████████████████████████▎         | 899/1230 [6:48:00<2:28:41, 26.95s/it]
                                                                                
{'loss': 0.6615, 'grad_norm': 0.39644289016723633, 'learning_rate': 4.6988644227199424e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.78, 'epoch': 0.64}

 73%|██████████████████████████▎         | 899/1230 [6:48:00<2:28:41, 26.95s/it]
 73%|██████████████████████████▎         | 900/1230 [6:48:30<2:32:37, 27.75s/it]
                                                                                
{'loss': 0.6789, 'grad_norm': 0.3867088556289673, 'learning_rate': 4.6825442583216684e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.63, 'epoch': 0.64}

 73%|██████████████████████████▎         | 900/1230 [6:48:30<2:32:37, 27.75s/it]
 73%|██████████████████████████▎         | 901/1230 [6:48:56<2:28:57, 27.17s/it]
                                                                                
{'loss': 0.7416, 'grad_norm': 0.41429686546325684, 'learning_rate': 4.666260214079201e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.41, 'epoch': 0.64}

 73%|██████████████████████████▎         | 901/1230 [6:48:56<2:28:57, 27.17s/it]
 73%|██████████████████████████▍         | 902/1230 [6:49:25<2:32:08, 27.83s/it]
                                                                                
{'loss': 0.7046, 'grad_norm': 0.3889736235141754, 'learning_rate': 4.650012402726074e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.58, 'epoch': 0.64}

 73%|██████████████████████████▍         | 902/1230 [6:49:25<2:32:08, 27.83s/it]
 73%|██████████████████████████▍         | 903/1230 [6:49:54<2:33:47, 28.22s/it]
                                                                                
{'loss': 0.661, 'grad_norm': 0.3681470453739166, 'learning_rate': 4.633800936744977e-05, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.13, 'epoch': 0.64}

 73%|██████████████████████████▍         | 903/1230 [6:49:54<2:33:47, 28.22s/it]
 73%|██████████████████████████▍         | 904/1230 [6:50:19<2:28:11, 27.28s/it]
                                                                                
{'loss': 0.7098, 'grad_norm': 0.42150571942329407, 'learning_rate': 4.6176259283669885e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.95, 'epoch': 0.64}

 73%|██████████████████████████▍         | 904/1230 [6:50:19<2:28:11, 27.28s/it]
 74%|██████████████████████████▍         | 905/1230 [6:50:45<2:25:04, 26.78s/it]
                                                                                
{'loss': 0.7305, 'grad_norm': 0.4091438055038452, 'learning_rate': 4.601487489570796e-05, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.61, 'epoch': 0.64}

 74%|██████████████████████████▍         | 905/1230 [6:50:45<2:25:04, 26.78s/it]
 74%|██████████████████████████▌         | 906/1230 [6:51:14<2:28:24, 27.48s/it]
                                                                                
{'loss': 0.6947, 'grad_norm': 0.37736690044403076, 'learning_rate': 4.585385732081907e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.07, 'epoch': 0.64}

 74%|██████████████████████████▌         | 906/1230 [6:51:14<2:28:24, 27.48s/it]
 74%|██████████████████████████▌         | 907/1230 [6:51:44<2:31:23, 28.12s/it]
                                                                                
{'loss': 0.7006, 'grad_norm': 0.3989769518375397, 'learning_rate': 4.5693207673718965e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.08, 'epoch': 0.64}

 74%|██████████████████████████▌         | 907/1230 [6:51:44<2:31:23, 28.12s/it]
 74%|██████████████████████████▌         | 908/1230 [6:52:13<2:33:35, 28.62s/it]
                                                                                
{'loss': 0.6921, 'grad_norm': 0.4287826120853424, 'learning_rate': 4.553292706657622e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.74, 'epoch': 0.64}

 74%|██████████████████████████▌         | 908/1230 [6:52:13<2:33:35, 28.62s/it]
 74%|██████████████████████████▌         | 909/1230 [6:52:43<2:34:41, 28.91s/it]
                                                                                
{'loss': 0.7144, 'grad_norm': 0.41485846042633057, 'learning_rate': 4.537301660900452e-05, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 144.38, 'epoch': 0.64}

 74%|██████████████████████████▌         | 909/1230 [6:52:43<2:34:41, 28.91s/it]
 74%|██████████████████████████▋         | 910/1230 [6:53:12<2:34:50, 29.03s/it]
                                                                                
{'loss': 0.6938, 'grad_norm': 0.40316644310951233, 'learning_rate': 4.521347740805509e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.36, 'epoch': 0.64}

 74%|██████████████████████████▋         | 910/1230 [6:53:12<2:34:50, 29.03s/it]
 74%|██████████████████████████▋         | 911/1230 [6:53:38<2:28:55, 28.01s/it]
                                                                                
{'loss': 0.7043, 'grad_norm': 0.41270366311073303, 'learning_rate': 4.505431056820894e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.8, 'epoch': 0.64}

 74%|██████████████████████████▋         | 911/1230 [6:53:38<2:28:55, 28.01s/it]
 74%|██████████████████████████▋         | 912/1230 [6:54:07<2:30:17, 28.36s/it]
                                                                                
{'loss': 0.7077, 'grad_norm': 0.4094064235687256, 'learning_rate': 4.489551719136925e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.99, 'epoch': 0.65}

 74%|██████████████████████████▋         | 912/1230 [6:54:07<2:30:17, 28.36s/it]
 74%|██████████████████████████▋         | 913/1230 [6:54:37<2:32:00, 28.77s/it]
                                                                                
{'loss': 0.6898, 'grad_norm': 0.4001717269420624, 'learning_rate': 4.473709837685375e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.11, 'epoch': 0.65}

 74%|██████████████████████████▋         | 913/1230 [6:54:37<2:32:00, 28.77s/it]
 74%|██████████████████████████▊         | 914/1230 [6:55:03<2:26:51, 27.88s/it]
                                                                                
{'loss': 0.6709, 'grad_norm': 0.3729688823223114, 'learning_rate': 4.457905522138705e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.27, 'epoch': 0.65}

 74%|██████████████████████████▊         | 914/1230 [6:55:03<2:26:51, 27.88s/it]
 74%|██████████████████████████▊         | 915/1230 [6:55:28<2:22:50, 27.21s/it]
                                                                                
{'loss': 0.6835, 'grad_norm': 0.3863973021507263, 'learning_rate': 4.442138881909311e-05, 'memory/max_active (GiB)': 9.88, 'memory/max_allocated (GiB)': 9.88, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 202.22, 'epoch': 0.65}

 74%|██████████████████████████▊         | 915/1230 [6:55:28<2:22:50, 27.21s/it]
 74%|██████████████████████████▊         | 916/1230 [6:55:57<2:25:28, 27.80s/it]
                                                                                
{'loss': 0.7138, 'grad_norm': 0.40139448642730713, 'learning_rate': 4.42641002614877e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.71, 'epoch': 0.65}

 74%|██████████████████████████▊         | 916/1230 [6:55:58<2:25:28, 27.80s/it]
 75%|██████████████████████████▊         | 917/1230 [6:56:23<2:21:11, 27.06s/it]
                                                                                
{'loss': 0.6593, 'grad_norm': 0.39192837476730347, 'learning_rate': 4.410719063747073e-05, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.98, 'epoch': 0.65}

 75%|██████████████████████████▊         | 917/1230 [6:56:23<2:21:11, 27.06s/it]
 75%|██████████████████████████▊         | 918/1230 [6:56:53<2:24:49, 27.85s/it]
                                                                                
{'loss': 0.7264, 'grad_norm': 0.42110615968704224, 'learning_rate': 4.395066103331882e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.43, 'epoch': 0.65}

 75%|██████████████████████████▊         | 918/1230 [6:56:53<2:24:49, 27.85s/it]
 75%|██████████████████████████▉         | 919/1230 [6:57:22<2:27:16, 28.41s/it]
                                                                                
{'loss': 0.6996, 'grad_norm': 0.3938377797603607, 'learning_rate': 4.379451253267768e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.89, 'epoch': 0.65}

 75%|██████████████████████████▉         | 919/1230 [6:57:22<2:27:16, 28.41s/it]
 75%|██████████████████████████▉         | 920/1230 [6:57:48<2:22:04, 27.50s/it]
                                                                                
{'loss': 0.6778, 'grad_norm': 0.39774712920188904, 'learning_rate': 4.363874621655469e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.5, 'epoch': 0.65}

 75%|██████████████████████████▉         | 920/1230 [6:57:48<2:22:04, 27.50s/it]
 75%|██████████████████████████▉         | 921/1230 [6:58:17<2:24:53, 28.13s/it]
                                                                                
{'loss': 0.7223, 'grad_norm': 0.39001697301864624, 'learning_rate': 4.348336316331141e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.21, 'epoch': 0.65}

 75%|██████████████████████████▉         | 921/1230 [6:58:17<2:24:53, 28.13s/it]
 75%|██████████████████████████▉         | 922/1230 [6:58:47<2:26:27, 28.53s/it]
                                                                                
{'loss': 0.6852, 'grad_norm': 0.3865410387516022, 'learning_rate': 4.332836444865607e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.54, 'epoch': 0.65}

 75%|██████████████████████████▉         | 922/1230 [6:58:47<2:26:27, 28.53s/it]
 75%|███████████████████████████         | 923/1230 [6:59:16<2:27:21, 28.80s/it]
                                                                                
{'loss': 0.7108, 'grad_norm': 0.40618860721588135, 'learning_rate': 4.317375114563617e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.98, 'epoch': 0.65}

 75%|███████████████████████████         | 923/1230 [6:59:16<2:27:21, 28.80s/it]
 75%|███████████████████████████         | 924/1230 [6:59:46<2:27:46, 28.98s/it]
                                                                                
{'loss': 0.6921, 'grad_norm': 0.38798990845680237, 'learning_rate': 4.301952432463094e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.69, 'epoch': 0.65}

 75%|███████████████████████████         | 924/1230 [6:59:46<2:27:46, 28.98s/it]
 75%|███████████████████████████         | 925/1230 [7:00:11<2:22:20, 28.00s/it]
                                                                                
{'loss': 0.6793, 'grad_norm': 0.3968304395675659, 'learning_rate': 4.2865685053344134e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.28, 'epoch': 0.65}

 75%|███████████████████████████         | 925/1230 [7:00:11<2:22:20, 28.00s/it]
 75%|███████████████████████████         | 926/1230 [7:00:40<2:23:29, 28.32s/it]
                                                                                
{'loss': 0.7339, 'grad_norm': 0.4064173400402069, 'learning_rate': 4.271223439679645e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.7, 'epoch': 0.66}

 75%|███████████████████████████         | 926/1230 [7:00:40<2:23:29, 28.32s/it]
 75%|███████████████████████████▏        | 927/1230 [7:01:05<2:18:15, 27.38s/it]
                                                                                
{'loss': 0.6818, 'grad_norm': 0.4268450438976288, 'learning_rate': 4.2559173417318236e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.22, 'epoch': 0.66}

 75%|███████████████████████████▏        | 927/1230 [7:01:05<2:18:15, 27.38s/it]
 75%|███████████████████████████▏        | 928/1230 [7:01:31<2:14:34, 26.74s/it]
                                                                                
{'loss': 0.6848, 'grad_norm': 0.41271114349365234, 'learning_rate': 4.240650317454218e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.02, 'epoch': 0.66}

 75%|███████████████████████████▏        | 928/1230 [7:01:31<2:14:34, 26.74s/it]
 76%|███████████████████████████▏        | 929/1230 [7:01:56<2:11:36, 26.24s/it]
                                                                                
{'loss': 0.6646, 'grad_norm': 0.40417295694351196, 'learning_rate': 4.225422472539582e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.95, 'epoch': 0.66}

 76%|███████████████████████████▏        | 929/1230 [7:01:56<2:11:36, 26.24s/it]
 76%|███████████████████████████▏        | 930/1230 [7:02:22<2:10:56, 26.19s/it]
                                                                                
{'loss': 0.7246, 'grad_norm': 0.4191085398197174, 'learning_rate': 4.210233912409443e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.36, 'epoch': 0.66}

 76%|███████████████████████████▏        | 930/1230 [7:02:22<2:10:56, 26.19s/it]
 76%|███████████████████████████▏        | 931/1230 [7:02:47<2:08:52, 25.86s/it]
                                                                                
{'loss': 0.6972, 'grad_norm': 0.39517879486083984, 'learning_rate': 4.1950847422133554e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.5, 'epoch': 0.66}

 76%|███████████████████████████▏        | 931/1230 [7:02:47<2:08:52, 25.86s/it]
 76%|███████████████████████████▎        | 932/1230 [7:03:12<2:07:15, 25.62s/it]
                                                                                
{'loss': 0.7109, 'grad_norm': 0.4225514829158783, 'learning_rate': 4.179975066828184e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.01, 'epoch': 0.66}

 76%|███████████████████████████▎        | 932/1230 [7:03:12<2:07:15, 25.62s/it]
 76%|███████████████████████████▎        | 933/1230 [7:03:37<2:06:09, 25.49s/it]
                                                                                
{'loss': 0.6495, 'grad_norm': 0.38590025901794434, 'learning_rate': 4.1649049908573666e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.87, 'epoch': 0.66}

 76%|███████████████████████████▎        | 933/1230 [7:03:37<2:06:09, 25.49s/it]
 76%|███████████████████████████▎        | 934/1230 [7:04:03<2:06:18, 25.60s/it]
                                                                                
{'loss': 0.6678, 'grad_norm': 0.3824492394924164, 'learning_rate': 4.149874618630206e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.03, 'epoch': 0.66}

 76%|███████████████████████████▎        | 934/1230 [7:04:03<2:06:18, 25.60s/it]
 76%|███████████████████████████▎        | 935/1230 [7:04:28<2:05:13, 25.47s/it]
                                                                                
{'loss': 0.6756, 'grad_norm': 0.405088871717453, 'learning_rate': 4.1348840542011276e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.06, 'epoch': 0.66}

 76%|███████████████████████████▎        | 935/1230 [7:04:28<2:05:13, 25.47s/it]
 76%|███████████████████████████▍        | 936/1230 [7:04:58<2:10:48, 26.70s/it]
                                                                                
{'loss': 0.6555, 'grad_norm': 0.38152021169662476, 'learning_rate': 4.1199334013489776e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.55, 'epoch': 0.66}

 76%|███████████████████████████▍        | 936/1230 [7:04:58<2:10:48, 26.70s/it]
 76%|███████████████████████████▍        | 937/1230 [7:05:28<2:14:48, 27.61s/it]
                                                                                
{'loss': 0.6924, 'grad_norm': 0.3978879153728485, 'learning_rate': 4.1050227635762936e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.17, 'epoch': 0.66}

 76%|███████████████████████████▍        | 937/1230 [7:05:28<2:14:48, 27.61s/it]
 76%|███████████████████████████▍        | 938/1230 [7:05:57<2:17:28, 28.25s/it]
                                                                                
{'loss': 0.6742, 'grad_norm': 0.39678677916526794, 'learning_rate': 4.090152244108593e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.2, 'epoch': 0.66}

 76%|███████████████████████████▍        | 938/1230 [7:05:57<2:17:28, 28.25s/it]
 76%|███████████████████████████▍        | 939/1230 [7:06:27<2:19:20, 28.73s/it]
                                                                                
{'loss': 0.6922, 'grad_norm': 0.4422370493412018, 'learning_rate': 4.075321945893651e-05, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.05, 'epoch': 0.66}

 76%|███████████████████████████▍        | 939/1230 [7:06:27<2:19:20, 28.73s/it]
 76%|███████████████████████████▌        | 940/1230 [7:06:52<2:14:00, 27.72s/it]
                                                                                
{'loss': 0.7075, 'grad_norm': 0.4174067974090576, 'learning_rate': 4.060531971600799e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.79, 'epoch': 0.67}

 76%|███████████████████████████▌        | 940/1230 [7:06:53<2:14:00, 27.72s/it]
 77%|███████████████████████████▌        | 941/1230 [7:07:18<2:10:27, 27.08s/it]
                                                                                
{'loss': 0.6727, 'grad_norm': 0.4013630151748657, 'learning_rate': 4.045782423620206e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.5, 'epoch': 0.67}

 77%|███████████████████████████▌        | 941/1230 [7:07:18<2:10:27, 27.08s/it]
 77%|███████████████████████████▌        | 942/1230 [7:07:48<2:14:00, 27.92s/it]
                                                                                
{'loss': 0.6973, 'grad_norm': 0.40200114250183105, 'learning_rate': 4.031073404062172e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.58, 'epoch': 0.67}

 77%|███████████████████████████▌        | 942/1230 [7:07:48<2:14:00, 27.92s/it]
 77%|███████████████████████████▌        | 943/1230 [7:08:14<2:10:38, 27.31s/it]
                                                                                
{'loss': 0.7125, 'grad_norm': 0.4094705879688263, 'learning_rate': 4.016405014756424e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.74, 'epoch': 0.67}

 77%|███████████████████████████▌        | 943/1230 [7:08:14<2:10:38, 27.31s/it]
 77%|███████████████████████████▋        | 944/1230 [7:08:43<2:12:57, 27.89s/it]
                                                                                
{'loss': 0.7114, 'grad_norm': 0.3936178982257843, 'learning_rate': 4.001777357251399e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.3, 'epoch': 0.67}

 77%|███████████████████████████▋        | 944/1230 [7:08:43<2:12:57, 27.89s/it]
 77%|███████████████████████████▋        | 945/1230 [7:09:08<2:08:46, 27.11s/it]
                                                                                
{'loss': 0.6823, 'grad_norm': 0.4030247926712036, 'learning_rate': 3.987190532813559e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.87, 'epoch': 0.67}

 77%|███████████████████████████▋        | 945/1230 [7:09:08<2:08:46, 27.11s/it]
 77%|███████████████████████████▋        | 946/1230 [7:09:38<2:11:28, 27.78s/it]
                                                                                
{'loss': 0.7328, 'grad_norm': 0.40389284491539, 'learning_rate': 3.972644642426679e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.11, 'epoch': 0.67}

 77%|███████████████████████████▋        | 946/1230 [7:09:38<2:11:28, 27.78s/it]
 77%|███████████████████████████▋        | 947/1230 [7:10:07<2:13:00, 28.20s/it]
                                                                                
{'loss': 0.7127, 'grad_norm': 0.39842528104782104, 'learning_rate': 3.958139786791147e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.3, 'epoch': 0.67}

 77%|███████████████████████████▋        | 947/1230 [7:10:07<2:13:00, 28.20s/it]
 77%|███████████████████████████▋        | 948/1230 [7:10:32<2:08:25, 27.32s/it]
                                                                                
{'loss': 0.6965, 'grad_norm': 0.41700735688209534, 'learning_rate': 3.9436760663232766e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.8, 'epoch': 0.67}

 77%|███████████████████████████▋        | 948/1230 [7:10:32<2:08:25, 27.32s/it]
 77%|███████████████████████████▊        | 949/1230 [7:10:58<2:06:02, 26.91s/it]
                                                                                
{'loss': 0.6678, 'grad_norm': 0.37824293971061707, 'learning_rate': 3.929253581154592e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.95, 'epoch': 0.67}

 77%|███████████████████████████▊        | 949/1230 [7:10:58<2:06:02, 26.91s/it]
 77%|███████████████████████████▊        | 950/1230 [7:11:28<2:09:10, 27.68s/it]
                                                                                
{'loss': 0.7067, 'grad_norm': 0.4229357838630676, 'learning_rate': 3.914872431131159e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.9, 'epoch': 0.67}

 77%|███████████████████████████▊        | 950/1230 [7:11:28<2:09:10, 27.68s/it]
 77%|███████████████████████████▊        | 951/1230 [7:11:57<2:11:04, 28.19s/it]
                                                                                
{'loss': 0.7247, 'grad_norm': 0.4200492203235626, 'learning_rate': 3.900532715812881e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.66, 'epoch': 0.67}

 77%|███████████████████████████▊        | 951/1230 [7:11:57<2:11:04, 28.19s/it]
 77%|███████████████████████████▊        | 952/1230 [7:12:26<2:12:26, 28.59s/it]
                                                                                
{'loss': 0.682, 'grad_norm': 0.4018292725086212, 'learning_rate': 3.886234534472806e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.62, 'epoch': 0.67}

 77%|███████████████████████████▊        | 952/1230 [7:12:27<2:12:26, 28.59s/it]
 77%|███████████████████████████▉        | 953/1230 [7:12:55<2:12:26, 28.69s/it]
                                                                                
{'loss': 0.6964, 'grad_norm': 0.41436296701431274, 'learning_rate': 3.8719779860964536e-05, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.96, 'epoch': 0.67}

 77%|███████████████████████████▉        | 953/1230 [7:12:55<2:12:26, 28.69s/it]
 78%|███████████████████████████▉        | 954/1230 [7:13:25<2:12:47, 28.87s/it]
                                                                                
{'loss': 0.6914, 'grad_norm': 0.4158988893032074, 'learning_rate': 3.857763169381108e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.15, 'epoch': 0.68}

 78%|███████████████████████████▉        | 954/1230 [7:13:25<2:12:47, 28.87s/it]
 78%|███████████████████████████▉        | 955/1230 [7:13:50<2:07:49, 27.89s/it]
                                                                                
{'loss': 0.689, 'grad_norm': 0.4028582274913788, 'learning_rate': 3.843590182735157e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.07, 'epoch': 0.68}

 78%|███████████████████████████▉        | 955/1230 [7:13:50<2:07:49, 27.89s/it]
 78%|███████████████████████████▉        | 956/1230 [7:14:20<2:09:32, 28.37s/it]
                                                                                
{'loss': 0.6852, 'grad_norm': 0.415134459733963, 'learning_rate': 3.829459124277399e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.17, 'epoch': 0.68}

 78%|███████████████████████████▉        | 956/1230 [7:14:20<2:09:32, 28.37s/it]
 78%|████████████████████████████        | 957/1230 [7:14:45<2:05:22, 27.55s/it]
                                                                                
{'loss': 0.6808, 'grad_norm': 0.39957454800605774, 'learning_rate': 3.815370091836366e-05, 'memory/max_active (GiB)': 10.38, 'memory/max_allocated (GiB)': 10.38, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.36, 'epoch': 0.68}

 78%|████████████████████████████        | 957/1230 [7:14:45<2:05:22, 27.55s/it]
 78%|████████████████████████████        | 958/1230 [7:15:15<2:07:09, 28.05s/it]
                                                                                
{'loss': 0.6871, 'grad_norm': 0.39767858386039734, 'learning_rate': 3.8013231829496474e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.95, 'epoch': 0.68}

 78%|████████████████████████████        | 958/1230 [7:15:15<2:07:09, 28.05s/it]
 78%|████████████████████████████        | 959/1230 [7:15:44<2:08:29, 28.45s/it]
                                                                                
{'loss': 0.6959, 'grad_norm': 0.4262488782405853, 'learning_rate': 3.787318494863207e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.55, 'epoch': 0.68}

 78%|████████████████████████████        | 959/1230 [7:15:44<2:08:29, 28.45s/it]
 78%|████████████████████████████        | 960/1230 [7:16:13<2:08:43, 28.61s/it]
                                                                                
{'loss': 0.6553, 'grad_norm': 0.390865683555603, 'learning_rate': 3.7733561245307234e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.82, 'epoch': 0.68}

 78%|████████████████████████████        | 960/1230 [7:16:13<2:08:43, 28.61s/it]
 78%|████████████████████████████▏       | 961/1230 [7:16:39<2:04:25, 27.75s/it]
                                                                                
{'loss': 0.6904, 'grad_norm': 0.3953944146633148, 'learning_rate': 3.759436168612913e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.21, 'epoch': 0.68}

 78%|████████████████████████████▏       | 961/1230 [7:16:39<2:04:25, 27.75s/it]
 78%|████████████████████████████▏       | 962/1230 [7:17:05<2:01:51, 27.28s/it]
                                                                                
{'loss': 0.6926, 'grad_norm': 0.3993051052093506, 'learning_rate': 3.745558723476851e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.22, 'epoch': 0.68}

 78%|████████████████████████████▏       | 962/1230 [7:17:05<2:01:51, 27.28s/it]
 78%|████████████████████████████▏       | 963/1230 [7:17:31<1:59:45, 26.91s/it]
                                                                                
{'loss': 0.7171, 'grad_norm': 0.40074455738067627, 'learning_rate': 3.731723885195319e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.4, 'epoch': 0.68}

 78%|████████████████████████████▏       | 963/1230 [7:17:31<1:59:45, 26.91s/it]
 78%|████████████████████████████▏       | 964/1230 [7:18:01<2:02:59, 27.74s/it]
                                                                                
{'loss': 0.693, 'grad_norm': 0.38722091913223267, 'learning_rate': 3.717931749546137e-05, 'memory/max_active (GiB)': 10.26, 'memory/max_allocated (GiB)': 10.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.92, 'epoch': 0.68}

 78%|████████████████████████████▏       | 964/1230 [7:18:01<2:02:59, 27.74s/it]
 78%|████████████████████████████▏       | 965/1230 [7:18:30<2:05:11, 28.35s/it]
                                                                                
{'loss': 0.7365, 'grad_norm': 0.4097769260406494, 'learning_rate': 3.7041824120114877e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.33, 'epoch': 0.68}

 78%|████████████████████████████▏       | 965/1230 [7:18:30<2:05:11, 28.35s/it]
 79%|████████████████████████████▎       | 966/1230 [7:18:56<2:00:42, 27.43s/it]
                                                                                
{'loss': 0.6938, 'grad_norm': 0.391764760017395, 'learning_rate': 3.69047596777727e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.26, 'epoch': 0.68}

 79%|████████████████████████████▎       | 966/1230 [7:18:56<2:00:42, 27.43s/it]
 79%|████████████████████████████▎       | 967/1230 [7:19:25<2:03:13, 28.11s/it]
                                                                                
{'loss': 0.6694, 'grad_norm': 0.4152336120605469, 'learning_rate': 3.676812511732438e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.71, 'epoch': 0.68}

 79%|████████████████████████████▎       | 967/1230 [7:19:25<2:03:13, 28.11s/it]
 79%|████████████████████████████▎       | 968/1230 [7:19:51<1:59:17, 27.32s/it]
                                                                                
{'loss': 0.6566, 'grad_norm': 0.4121333360671997, 'learning_rate': 3.6631921384683376e-05, 'memory/max_active (GiB)': 11.29, 'memory/max_allocated (GiB)': 11.29, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.95, 'epoch': 0.69}

 79%|████████████████████████████▎       | 968/1230 [7:19:51<1:59:17, 27.32s/it]
 79%|████████████████████████████▎       | 969/1230 [7:20:20<2:01:18, 27.89s/it]
                                                                                
{'loss': 0.7088, 'grad_norm': 0.38556262850761414, 'learning_rate': 3.64961494227805e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.54, 'epoch': 0.69}

 79%|████████████████████████████▎       | 969/1230 [7:20:20<2:01:18, 27.89s/it]
 79%|████████████████████████████▍       | 970/1230 [7:20:45<1:56:55, 26.98s/it]
                                                                                
{'loss': 0.6708, 'grad_norm': 0.3926299512386322, 'learning_rate': 3.636081017155755e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.25, 'epoch': 0.69}

 79%|████████████████████████████▍       | 970/1230 [7:20:45<1:56:55, 26.98s/it]
 79%|████████████████████████████▍       | 971/1230 [7:21:14<1:58:52, 27.54s/it]
                                                                                
{'loss': 0.6888, 'grad_norm': 0.39231351017951965, 'learning_rate': 3.6225904567960595e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.71, 'epoch': 0.69}

 79%|████████████████████████████▍       | 971/1230 [7:21:14<1:58:52, 27.54s/it]
 79%|████████████████████████████▍       | 972/1230 [7:21:39<1:55:47, 26.93s/it]
                                                                                
{'loss': 0.6727, 'grad_norm': 0.40444216132164, 'learning_rate': 3.609143354593363e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.3, 'epoch': 0.69}

 79%|████████████████████████████▍       | 972/1230 [7:21:39<1:55:47, 26.93s/it]
 79%|████████████████████████████▍       | 973/1230 [7:22:05<1:53:45, 26.56s/it]
                                                                                
{'loss': 0.6838, 'grad_norm': 0.3984462320804596, 'learning_rate': 3.59573980364121e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.07, 'epoch': 0.69}

 79%|████████████████████████████▍       | 973/1230 [7:22:05<1:53:45, 26.56s/it]
 79%|████████████████████████████▌       | 974/1230 [7:22:31<1:52:08, 26.28s/it]
                                                                                
{'loss': 0.6735, 'grad_norm': 0.3943852186203003, 'learning_rate': 3.582379896731631e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.81, 'epoch': 0.69}

 79%|████████████████████████████▌       | 974/1230 [7:22:31<1:52:08, 26.28s/it]
 79%|████████████████████████████▌       | 975/1230 [7:23:00<1:55:25, 27.16s/it]
                                                                                
{'loss': 0.6985, 'grad_norm': 0.3798525929450989, 'learning_rate': 3.5690637263545216e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.64, 'epoch': 0.69}

 79%|████████████████████████████▌       | 975/1230 [7:23:00<1:55:25, 27.16s/it]
 79%|████████████████████████████▌       | 976/1230 [7:23:29<1:57:38, 27.79s/it]
                                                                                
{'loss': 0.6724, 'grad_norm': 0.3742969334125519, 'learning_rate': 3.5557913846969877e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.25, 'epoch': 0.69}

 79%|████████████████████████████▌       | 976/1230 [7:23:29<1:57:38, 27.79s/it]
 79%|████████████████████████████▌       | 977/1230 [7:23:58<1:58:54, 28.20s/it]
                                                                                
{'loss': 0.6765, 'grad_norm': 0.41923534870147705, 'learning_rate': 3.542562963642714e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.32, 'epoch': 0.69}

 79%|████████████████████████████▌       | 977/1230 [7:23:58<1:58:54, 28.20s/it]
 80%|████████████████████████████▌       | 978/1230 [7:24:28<2:00:22, 28.66s/it]
                                                                                
{'loss': 0.6793, 'grad_norm': 0.41756898164749146, 'learning_rate': 3.5293785547713245e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.09, 'epoch': 0.69}

 80%|████████████████████████████▌       | 978/1230 [7:24:28<2:00:22, 28.66s/it]
 80%|████████████████████████████▋       | 979/1230 [7:24:54<1:56:13, 27.78s/it]
                                                                                
{'loss': 0.69, 'grad_norm': 0.398899108171463, 'learning_rate': 3.516238249357743e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.93, 'epoch': 0.69}

 80%|████████████████████████████▋       | 979/1230 [7:24:54<1:56:13, 27.78s/it]
 80%|████████████████████████████▋       | 980/1230 [7:25:20<1:53:15, 27.18s/it]
                                                                                
{'loss': 0.6856, 'grad_norm': 0.4092264771461487, 'learning_rate': 3.503142138371576e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.18, 'epoch': 0.69}

 80%|████████████████████████████▋       | 980/1230 [7:25:20<1:53:15, 27.18s/it]
 80%|████████████████████████████▋       | 981/1230 [7:25:45<1:50:30, 26.63s/it]
                                                                                
{'loss': 0.661, 'grad_norm': 0.43092891573905945, 'learning_rate': 3.490090312476471e-05, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.79, 'epoch': 0.69}

 80%|████████████████████████████▋       | 981/1230 [7:25:45<1:50:30, 26.63s/it]
 80%|████████████████████████████▋       | 982/1230 [7:26:11<1:49:23, 26.47s/it]
                                                                                
{'loss': 0.7023, 'grad_norm': 0.40968507528305054, 'learning_rate': 3.477082862029493e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.49, 'epoch': 0.69}

 80%|████████████████████████████▋       | 982/1230 [7:26:11<1:49:23, 26.47s/it]
 80%|████████████████████████████▊       | 983/1230 [7:26:40<1:52:37, 27.36s/it]
                                                                                
{'loss': 0.6876, 'grad_norm': 0.4012763798236847, 'learning_rate': 3.464119877080499e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.28, 'epoch': 0.7}

 80%|████████████████████████████▊       | 983/1230 [7:26:40<1:52:37, 27.36s/it]
 80%|████████████████████████████▊       | 984/1230 [7:27:07<1:50:52, 27.04s/it]
                                                                                
{'loss': 0.6952, 'grad_norm': 0.41252920031547546, 'learning_rate': 3.4512014473715084e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.25, 'epoch': 0.7}

 80%|████████████████████████████▊       | 984/1230 [7:27:07<1:50:52, 27.04s/it]
 80%|████████████████████████████▊       | 985/1230 [7:27:33<1:49:01, 26.70s/it]
                                                                                
{'loss': 0.7033, 'grad_norm': 0.40046703815460205, 'learning_rate': 3.438327662336094e-05, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.0, 'epoch': 0.7}

 80%|████████████████████████████▊       | 985/1230 [7:27:33<1:49:01, 26.70s/it]
 80%|████████████████████████████▊       | 986/1230 [7:28:02<1:51:40, 27.46s/it]
                                                                                
{'loss': 0.7173, 'grad_norm': 0.39509493112564087, 'learning_rate': 3.4254986110987525e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.81, 'epoch': 0.7}

 80%|████████████████████████████▊       | 986/1230 [7:28:02<1:51:40, 27.46s/it]
 80%|████████████████████████████▉       | 987/1230 [7:28:27<1:48:43, 26.85s/it]
                                                                                
{'loss': 0.6941, 'grad_norm': 0.40939193964004517, 'learning_rate': 3.412714382474292e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.42, 'epoch': 0.7}

 80%|████████████████████████████▉       | 987/1230 [7:28:27<1:48:43, 26.85s/it]
 80%|████████████████████████████▉       | 988/1230 [7:28:57<1:51:17, 27.59s/it]
                                                                                
{'loss': 0.6938, 'grad_norm': 0.39736422896385193, 'learning_rate': 3.399975064967219e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.58, 'epoch': 0.7}

 80%|████████████████████████████▉       | 988/1230 [7:28:57<1:51:17, 27.59s/it]
 80%|████████████████████████████▉       | 989/1230 [7:29:22<1:48:11, 26.93s/it]
                                                                                
{'loss': 0.6947, 'grad_norm': 0.42991170287132263, 'learning_rate': 3.3872807467711154e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.66, 'epoch': 0.7}

 80%|████████████████████████████▉       | 989/1230 [7:29:22<1:48:11, 26.93s/it]
 80%|████████████████████████████▉       | 990/1230 [7:29:47<1:45:52, 26.47s/it]
                                                                                
{'loss': 0.6588, 'grad_norm': 0.3995579183101654, 'learning_rate': 3.374631515768043e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.21, 'epoch': 0.7}

 80%|████████████████████████████▉       | 990/1230 [7:29:47<1:45:52, 26.47s/it]
 81%|█████████████████████████████       | 991/1230 [7:30:17<1:49:19, 27.45s/it]
                                                                                
{'loss': 0.7155, 'grad_norm': 0.40563446283340454, 'learning_rate': 3.36202745952792e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.65, 'epoch': 0.7}

 81%|█████████████████████████████       | 991/1230 [7:30:17<1:49:19, 27.45s/it]
 81%|█████████████████████████████       | 992/1230 [7:30:42<1:46:21, 26.81s/it]
                                                                                
{'loss': 0.6891, 'grad_norm': 0.41971486806869507, 'learning_rate': 3.34946866530793e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.9, 'epoch': 0.7}

 81%|█████████████████████████████       | 992/1230 [7:30:42<1:46:21, 26.81s/it]
 81%|█████████████████████████████       | 993/1230 [7:31:08<1:44:03, 26.34s/it]
                                                                                
{'loss': 0.656, 'grad_norm': 0.41800978779792786, 'learning_rate': 3.3369552200519074e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.4, 'epoch': 0.7}

 81%|█████████████████████████████       | 993/1230 [7:31:08<1:44:03, 26.34s/it]
 81%|█████████████████████████████       | 994/1230 [7:31:33<1:42:56, 26.17s/it]
                                                                                
{'loss': 0.7028, 'grad_norm': 0.3962210416793823, 'learning_rate': 3.324487210389734e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.29, 'epoch': 0.7}

 81%|█████████████████████████████       | 994/1230 [7:31:33<1:42:56, 26.17s/it]
 81%|█████████████████████████████       | 995/1230 [7:32:03<1:46:02, 27.07s/it]
                                                                                
{'loss': 0.6944, 'grad_norm': 0.3967425227165222, 'learning_rate': 3.312064722636748e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.41, 'epoch': 0.7}

 81%|█████████████████████████████       | 995/1230 [7:32:03<1:46:02, 27.07s/it]
 81%|█████████████████████████████▏      | 996/1230 [7:32:32<1:48:25, 27.80s/it]
                                                                                
{'loss': 0.6937, 'grad_norm': 0.3954819440841675, 'learning_rate': 3.2996878427931435e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.95, 'epoch': 0.7}

 81%|█████████████████████████████▏      | 996/1230 [7:32:32<1:48:25, 27.80s/it]
 81%|█████████████████████████████▏      | 997/1230 [7:32:57<1:44:33, 26.93s/it]
                                                                                
{'loss': 0.7247, 'grad_norm': 0.41751283407211304, 'learning_rate': 3.287356656543368e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.74, 'epoch': 0.71}

 81%|█████████████████████████████▏      | 997/1230 [7:32:57<1:44:33, 26.93s/it]
 81%|█████████████████████████████▏      | 998/1230 [7:33:22<1:41:52, 26.35s/it]
                                                                                
{'loss': 0.65, 'grad_norm': 0.4281230568885803, 'learning_rate': 3.2750712492555416e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.05, 'epoch': 0.71}

 81%|█████████████████████████████▏      | 998/1230 [7:33:22<1:41:52, 26.35s/it]
 81%|█████████████████████████████▏      | 999/1230 [7:33:47<1:40:04, 25.99s/it]
                                                                                
{'loss': 0.6858, 'grad_norm': 0.4292943477630615, 'learning_rate': 3.26283170598085e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.33, 'epoch': 0.71}

 81%|█████████████████████████████▏      | 999/1230 [7:33:47<1:40:04, 25.99s/it]
 81%|████████████████████████████▍      | 1000/1230 [7:34:13<1:39:14, 25.89s/it]
                                                                                
{'loss': 0.7121, 'grad_norm': 0.41112038493156433, 'learning_rate': 3.250638111452974e-05, 'memory/max_active (GiB)': 10.01, 'memory/max_allocated (GiB)': 10.01, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 201.01, 'epoch': 0.71}

 81%|████████████████████████████▍      | 1000/1230 [7:34:13<1:39:14, 25.89s/it]
 81%|████████████████████████████▍      | 1001/1230 [7:34:42<1:42:55, 26.97s/it]
                                                                                
{'loss': 0.6908, 'grad_norm': 0.3888337016105652, 'learning_rate': 3.23849055008749e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.91, 'epoch': 0.71}

 81%|████████████████████████████▍      | 1001/1230 [7:34:42<1:42:55, 26.97s/it]
 81%|████████████████████████████▌      | 1002/1230 [7:35:12<1:45:38, 27.80s/it]
                                                                                
{'loss': 0.6812, 'grad_norm': 0.4204442799091339, 'learning_rate': 3.2263891059812876e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.54, 'epoch': 0.71}

 81%|████████████████████████████▌      | 1002/1230 [7:35:12<1:45:38, 27.80s/it]
 82%|████████████████████████████▌      | 1003/1230 [7:35:37<1:42:14, 27.02s/it]
                                                                                
{'loss': 0.6802, 'grad_norm': 0.41594794392585754, 'learning_rate': 3.214333862911995e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.79, 'epoch': 0.71}

 82%|████████████████████████████▌      | 1003/1230 [7:35:37<1:42:14, 27.02s/it]
 82%|████████████████████████████▌      | 1004/1230 [7:36:07<1:44:25, 27.72s/it]
                                                                                
{'loss': 0.6758, 'grad_norm': 0.37699374556541443, 'learning_rate': 3.202324904337383e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.6, 'epoch': 0.71}

 82%|████████████████████████████▌      | 1004/1230 [7:36:07<1:44:25, 27.72s/it]
 82%|████████████████████████████▌      | 1005/1230 [7:36:36<1:45:45, 28.20s/it]
                                                                                
{'loss': 0.7245, 'grad_norm': 0.4259316623210907, 'learning_rate': 3.190362313394808e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.77, 'epoch': 0.71}

 82%|████████████████████████████▌      | 1005/1230 [7:36:36<1:45:45, 28.20s/it]
 82%|████████████████████████████▋      | 1006/1230 [7:37:01<1:42:07, 27.36s/it]
                                                                                
{'loss': 0.668, 'grad_norm': 0.4278835356235504, 'learning_rate': 3.178446172900622e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.09, 'epoch': 0.71}

 82%|████████████████████████████▋      | 1006/1230 [7:37:01<1:42:07, 27.36s/it]
 82%|████████████████████████████▋      | 1007/1230 [7:37:27<1:39:49, 26.86s/it]
                                                                                
{'loss': 0.6781, 'grad_norm': 0.3925330340862274, 'learning_rate': 3.1665765653495983e-05, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 201.06, 'epoch': 0.71}

 82%|████████████████████████████▋      | 1007/1230 [7:37:27<1:39:49, 26.86s/it]
 82%|████████████████████████████▋      | 1008/1230 [7:37:56<1:42:05, 27.59s/it]
                                                                                
{'loss': 0.708, 'grad_norm': 0.412906676530838, 'learning_rate': 3.154753572914377e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.21, 'epoch': 0.71}

 82%|████████████████████████████▋      | 1008/1230 [7:37:56<1:42:05, 27.59s/it]
 82%|████████████████████████████▋      | 1009/1230 [7:38:23<1:40:18, 27.23s/it]
                                                                                
{'loss': 0.6978, 'grad_norm': 0.4307790994644165, 'learning_rate': 3.1429772774448694e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.37, 'epoch': 0.71}

 82%|████████████████████████████▋      | 1009/1230 [7:38:23<1:40:18, 27.23s/it]
 82%|████████████████████████████▋      | 1010/1230 [7:38:48<1:37:29, 26.59s/it]
                                                                                
{'loss': 0.6642, 'grad_norm': 0.4030582010746002, 'learning_rate': 3.131247760467716e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.41, 'epoch': 0.71}

 82%|████████████████████████████▋      | 1010/1230 [7:38:48<1:37:29, 26.59s/it]
 82%|████████████████████████████▊      | 1011/1230 [7:39:17<1:40:16, 27.47s/it]
                                                                                
{'loss': 0.7042, 'grad_norm': 0.38691386580467224, 'learning_rate': 3.119565103185712e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.36, 'epoch': 0.72}

 82%|████████████████████████████▊      | 1011/1230 [7:39:17<1:40:16, 27.47s/it]
 82%|████████████████████████████▊      | 1012/1230 [7:39:43<1:38:15, 27.04s/it]
                                                                                
{'loss': 0.6651, 'grad_norm': 0.3967471122741699, 'learning_rate': 3.1079293864772426e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.72, 'epoch': 0.72}

 82%|████████████████████████████▊      | 1012/1230 [7:39:43<1:38:15, 27.04s/it]
 82%|████████████████████████████▊      | 1013/1230 [7:40:08<1:35:35, 26.43s/it]
                                                                                
{'loss': 0.7309, 'grad_norm': 0.4174008369445801, 'learning_rate': 3.096340690895729e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.69, 'epoch': 0.72}

 82%|████████████████████████████▊      | 1013/1230 [7:40:08<1:35:35, 26.43s/it]
 82%|████████████████████████████▊      | 1014/1230 [7:40:38<1:38:28, 27.36s/it]
                                                                                
{'loss': 0.6447, 'grad_norm': 0.3893199563026428, 'learning_rate': 3.084799096669063e-05, 'memory/max_active (GiB)': 9.26, 'memory/max_allocated (GiB)': 9.26, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 148.42, 'epoch': 0.72}

 82%|████████████████████████████▊      | 1014/1230 [7:40:38<1:38:28, 27.36s/it]
 83%|████████████████████████████▉      | 1015/1230 [7:41:04<1:36:43, 26.99s/it]
                                                                                
{'loss': 0.6977, 'grad_norm': 0.4050295352935791, 'learning_rate': 3.0733046836990604e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.63, 'epoch': 0.72}

 83%|████████████████████████████▉      | 1015/1230 [7:41:04<1:36:43, 26.99s/it]
 83%|████████████████████████████▉      | 1016/1230 [7:41:29<1:34:07, 26.39s/it]
                                                                                
{'loss': 0.7008, 'grad_norm': 0.4114873707294464, 'learning_rate': 3.061857531560902e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.01, 'epoch': 0.72}

 83%|████████████████████████████▉      | 1016/1230 [7:41:29<1:34:07, 26.39s/it]
 83%|████████████████████████████▉      | 1017/1230 [7:41:58<1:36:34, 27.20s/it]
                                                                                
{'loss': 0.7086, 'grad_norm': 0.4289245903491974, 'learning_rate': 3.0504577195025883e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.25, 'epoch': 0.72}

 83%|████████████████████████████▉      | 1017/1230 [7:41:58<1:36:34, 27.20s/it]
 83%|████████████████████████████▉      | 1018/1230 [7:42:24<1:34:59, 26.89s/it]
                                                                                
{'loss': 0.6777, 'grad_norm': 0.38803035020828247, 'learning_rate': 3.0391053264443794e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.23, 'epoch': 0.72}

 83%|████████████████████████████▉      | 1018/1230 [7:42:24<1:34:59, 26.89s/it]
 83%|████████████████████████████▉      | 1019/1230 [7:42:50<1:32:50, 26.40s/it]
                                                                                
{'loss': 0.6936, 'grad_norm': 0.41363322734832764, 'learning_rate': 3.0278004309782637e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.16, 'epoch': 0.72}

 83%|████████████████████████████▉      | 1019/1230 [7:42:50<1:32:50, 26.40s/it]
 83%|█████████████████████████████      | 1020/1230 [7:43:19<1:35:28, 27.28s/it]
                                                                                
{'loss': 0.6724, 'grad_norm': 0.40045878291130066, 'learning_rate': 3.0165431113674036e-05, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.53, 'epoch': 0.72}

 83%|█████████████████████████████      | 1020/1230 [7:43:19<1:35:28, 27.28s/it]
 83%|█████████████████████████████      | 1021/1230 [7:43:45<1:33:54, 26.96s/it]
                                                                                
{'loss': 0.6894, 'grad_norm': 0.38745272159576416, 'learning_rate': 3.0053334455455943e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.56, 'epoch': 0.72}

 83%|█████████████████████████████      | 1021/1230 [7:43:45<1:33:54, 26.96s/it]
 83%|█████████████████████████████      | 1022/1230 [7:44:14<1:35:57, 27.68s/it]
                                                                                
{'loss': 0.6861, 'grad_norm': 0.3958342969417572, 'learning_rate': 2.99417151111673e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.8, 'epoch': 0.72}

 83%|█████████████████████████████      | 1022/1230 [7:44:14<1:35:57, 27.68s/it]
 83%|█████████████████████████████      | 1023/1230 [7:44:44<1:37:26, 28.24s/it]
                                                                                
{'loss': 0.6717, 'grad_norm': 0.39147546887397766, 'learning_rate': 2.9830573853542634e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.52, 'epoch': 0.72}

 83%|█████████████████████████████      | 1023/1230 [7:44:44<1:37:26, 28.24s/it]
 83%|█████████████████████████████▏     | 1024/1230 [7:45:13<1:38:07, 28.58s/it]
                                                                                
{'loss': 0.6694, 'grad_norm': 0.38278669118881226, 'learning_rate': 2.971991145200666e-05, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 147.39, 'epoch': 0.72}

 83%|█████████████████████████████▏     | 1024/1230 [7:45:13<1:38:07, 28.58s/it]
 83%|█████████████████████████████▏     | 1025/1230 [7:45:43<1:38:36, 28.86s/it]
                                                                                
{'loss': 0.7072, 'grad_norm': 0.40684813261032104, 'learning_rate': 2.9609728672669025e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.49, 'epoch': 0.73}

 83%|█████████████████████████████▏     | 1025/1230 [7:45:43<1:38:36, 28.86s/it]
 83%|█████████████████████████████▏     | 1026/1230 [7:46:08<1:34:09, 27.70s/it]
                                                                                
{'loss': 0.7118, 'grad_norm': 0.43260037899017334, 'learning_rate': 2.9500026278318987e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.27, 'epoch': 0.73}

 83%|█████████████████████████████▏     | 1026/1230 [7:46:08<1:34:09, 27.70s/it]
 83%|█████████████████████████████▏     | 1027/1230 [7:46:34<1:32:10, 27.24s/it]
                                                                                
{'loss': 0.6845, 'grad_norm': 0.4141291379928589, 'learning_rate': 2.939080502842013e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.71, 'epoch': 0.73}

 83%|█████████████████████████████▏     | 1027/1230 [7:46:34<1:32:10, 27.24s/it]
 84%|█████████████████████████████▎     | 1028/1230 [7:47:00<1:30:44, 26.95s/it]
                                                                                
{'loss': 0.7236, 'grad_norm': 0.4184190630912781, 'learning_rate': 2.92820656791051e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.17, 'epoch': 0.73}

 84%|█████████████████████████████▎     | 1028/1230 [7:47:00<1:30:44, 26.95s/it]
 84%|█████████████████████████████▎     | 1029/1230 [7:47:30<1:32:33, 27.63s/it]
                                                                                
{'loss': 0.6963, 'grad_norm': 0.4122909605503082, 'learning_rate': 2.9173808983170336e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.16, 'epoch': 0.73}

 84%|█████████████████████████████▎     | 1029/1230 [7:47:30<1:32:33, 27.63s/it]
 84%|█████████████████████████████▎     | 1030/1230 [7:47:55<1:29:39, 26.90s/it]
                                                                                
{'loss': 0.7277, 'grad_norm': 0.4577171206474304, 'learning_rate': 2.906603569007092e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.29, 'epoch': 0.73}

 84%|█████████████████████████████▎     | 1030/1230 [7:47:55<1:29:39, 26.90s/it]
 84%|█████████████████████████████▎     | 1031/1230 [7:48:20<1:27:41, 26.44s/it]
                                                                                
{'loss': 0.7112, 'grad_norm': 0.4140545427799225, 'learning_rate': 2.895874654591537e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.41, 'epoch': 0.73}

 84%|█████████████████████████████▎     | 1031/1230 [7:48:20<1:27:41, 26.44s/it]
 84%|█████████████████████████████▎     | 1032/1230 [7:48:50<1:30:19, 27.37s/it]
                                                                                
{'loss': 0.7274, 'grad_norm': 0.43488940596580505, 'learning_rate': 2.8851942293460466e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.59, 'epoch': 0.73}

 84%|█████████████████████████████▎     | 1032/1230 [7:48:50<1:30:19, 27.37s/it]
 84%|█████████████████████████████▍     | 1033/1230 [7:49:19<1:32:04, 28.04s/it]
                                                                                
{'loss': 0.7154, 'grad_norm': 0.4173978269100189, 'learning_rate': 2.8745623672106126e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.8, 'epoch': 0.73}

 84%|█████████████████████████████▍     | 1033/1230 [7:49:19<1:32:04, 28.04s/it]
 84%|█████████████████████████████▍     | 1034/1230 [7:49:45<1:29:03, 27.26s/it]
                                                                                
{'loss': 0.7091, 'grad_norm': 0.41404014825820923, 'learning_rate': 2.8639791417890208e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.45, 'epoch': 0.73}

 84%|█████████████████████████████▍     | 1034/1230 [7:49:45<1:29:03, 27.26s/it]
 84%|█████████████████████████████▍     | 1035/1230 [7:50:14<1:30:53, 27.97s/it]
                                                                                
{'loss': 0.6915, 'grad_norm': 0.438986599445343, 'learning_rate': 2.8534446263483545e-05, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.43, 'epoch': 0.73}

 84%|█████████████████████████████▍     | 1035/1230 [7:50:14<1:30:53, 27.97s/it]
 84%|█████████████████████████████▍     | 1036/1230 [7:50:40<1:27:49, 27.16s/it]
                                                                                
{'loss': 0.6784, 'grad_norm': 0.4025471806526184, 'learning_rate': 2.842958893818478e-05, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.8, 'epoch': 0.73}

 84%|█████████████████████████████▍     | 1036/1230 [7:50:40<1:27:49, 27.16s/it]
 84%|█████████████████████████████▌     | 1037/1230 [7:51:05<1:25:48, 26.68s/it]
                                                                                
{'loss': 0.6653, 'grad_norm': 0.41364622116088867, 'learning_rate': 2.8325220167915323e-05, 'memory/max_active (GiB)': 8.99, 'memory/max_allocated (GiB)': 8.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 210.07, 'epoch': 0.73}

 84%|█████████████████████████████▌     | 1037/1230 [7:51:05<1:25:48, 26.68s/it]
 84%|█████████████████████████████▌     | 1038/1230 [7:51:30<1:23:49, 26.20s/it]
                                                                                
{'loss': 0.687, 'grad_norm': 0.40253883600234985, 'learning_rate': 2.8221340675214404e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.93, 'epoch': 0.73}

 84%|█████████████████████████████▌     | 1038/1230 [7:51:30<1:23:49, 26.20s/it]
 84%|█████████████████████████████▌     | 1039/1230 [7:51:59<1:25:59, 27.01s/it]
                                                                                
{'loss': 0.6778, 'grad_norm': 0.3904304504394531, 'learning_rate': 2.811795117923392e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.88, 'epoch': 0.74}

 84%|█████████████████████████████▌     | 1039/1230 [7:51:59<1:25:59, 27.01s/it]
 85%|█████████████████████████████▌     | 1040/1230 [7:52:25<1:24:20, 26.63s/it]
                                                                                
{'loss': 0.6758, 'grad_norm': 0.3971782624721527, 'learning_rate': 2.8015052395733628e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.61, 'epoch': 0.74}

 85%|█████████████████████████████▌     | 1040/1230 [7:52:25<1:24:20, 26.63s/it]
 85%|█████████████████████████████▌     | 1041/1230 [7:52:50<1:22:34, 26.21s/it]
                                                                                
{'loss': 0.7166, 'grad_norm': 0.4459236264228821, 'learning_rate': 2.7912645037076066e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.01, 'epoch': 0.74}

 85%|█████████████████████████████▌     | 1041/1230 [7:52:50<1:22:34, 26.21s/it]
 85%|█████████████████████████████▋     | 1042/1230 [7:53:19<1:25:06, 27.16s/it]
                                                                                
{'loss': 0.6582, 'grad_norm': 0.3946848213672638, 'learning_rate': 2.7810729812221706e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.44, 'epoch': 0.74}

 85%|█████████████████████████████▋     | 1042/1230 [7:53:19<1:25:06, 27.16s/it]
 85%|█████████████████████████████▋     | 1043/1230 [7:53:49<1:27:08, 27.96s/it]
                                                                                
{'loss': 0.6518, 'grad_norm': 0.4014488160610199, 'learning_rate': 2.7709307426723972e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.52, 'epoch': 0.74}

 85%|█████████████████████████████▋     | 1043/1230 [7:53:49<1:27:08, 27.96s/it]
 85%|█████████████████████████████▋     | 1044/1230 [7:54:15<1:24:18, 27.20s/it]
                                                                                
{'loss': 0.6417, 'grad_norm': 0.3882829546928406, 'learning_rate': 2.760837858272438e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.73, 'epoch': 0.74}

 85%|█████████████████████████████▋     | 1044/1230 [7:54:15<1:24:18, 27.20s/it]
 85%|█████████████████████████████▋     | 1045/1230 [7:54:40<1:22:10, 26.65s/it]
                                                                                
{'loss': 0.6403, 'grad_norm': 0.41218554973602295, 'learning_rate': 2.75079439789477e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.33, 'epoch': 0.74}

 85%|█████████████████████████████▋     | 1045/1230 [7:54:40<1:22:10, 26.65s/it]
 85%|█████████████████████████████▊     | 1046/1230 [7:55:06<1:21:00, 26.41s/it]
                                                                                
{'loss': 0.7571, 'grad_norm': 0.4563050866127014, 'learning_rate': 2.740800431069714e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 198.0, 'epoch': 0.74}

 85%|█████████████████████████████▊     | 1046/1230 [7:55:06<1:21:00, 26.41s/it]
 85%|█████████████████████████████▊     | 1047/1230 [7:55:31<1:19:35, 26.10s/it]
                                                                                
{'loss': 0.6777, 'grad_norm': 0.3946678042411804, 'learning_rate': 2.730856026984941e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.52, 'epoch': 0.74}

 85%|█████████████████████████████▊     | 1047/1230 [7:55:31<1:19:35, 26.10s/it]
 85%|█████████████████████████████▊     | 1048/1230 [7:55:56<1:18:14, 25.79s/it]
                                                                                
{'loss': 0.6404, 'grad_norm': 0.4099472165107727, 'learning_rate': 2.720961254485011e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.31, 'epoch': 0.74}

 85%|█████████████████████████████▊     | 1048/1230 [7:55:56<1:18:14, 25.79s/it]
 85%|█████████████████████████████▊     | 1049/1230 [7:56:22<1:17:59, 25.86s/it]
                                                                                
{'loss': 0.696, 'grad_norm': 0.4042412340641022, 'learning_rate': 2.7111161820708836e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.11, 'epoch': 0.74}

 85%|█████████████████████████████▊     | 1049/1230 [7:56:22<1:17:59, 25.86s/it]
 85%|█████████████████████████████▉     | 1050/1230 [7:56:47<1:16:37, 25.54s/it]
                                                                                
{'loss': 0.6641, 'grad_norm': 0.4005156457424164, 'learning_rate': 2.7013208778994442e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.06, 'epoch': 0.74}

 85%|█████████████████████████████▉     | 1050/1230 [7:56:47<1:16:37, 25.54s/it]
 85%|█████████████████████████████▉     | 1051/1230 [7:57:16<1:19:17, 26.58s/it]
                                                                                
{'loss': 0.6755, 'grad_norm': 0.42772990465164185, 'learning_rate': 2.691575409783041e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.68, 'epoch': 0.74}

 85%|█████████████████████████████▉     | 1051/1230 [7:57:16<1:19:17, 26.58s/it]
 86%|█████████████████████████████▉     | 1052/1230 [7:57:41<1:17:30, 26.13s/it]
                                                                                
{'loss': 0.6841, 'grad_norm': 0.394910603761673, 'learning_rate': 2.6818798451890067e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.72, 'epoch': 0.74}

 86%|█████████████████████████████▉     | 1052/1230 [7:57:41<1:17:30, 26.13s/it]
 86%|█████████████████████████████▉     | 1053/1230 [7:58:06<1:16:09, 25.82s/it]
                                                                                
{'loss': 0.6888, 'grad_norm': 0.43729284405708313, 'learning_rate': 2.6722342512391983e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.27, 'epoch': 0.75}

 86%|█████████████████████████████▉     | 1053/1230 [7:58:06<1:16:09, 25.82s/it]
 86%|█████████████████████████████▉     | 1054/1230 [7:58:36<1:18:56, 26.91s/it]
                                                                                
{'loss': 0.6958, 'grad_norm': 0.4103975296020508, 'learning_rate': 2.662638694709521e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.23, 'epoch': 0.75}

 86%|█████████████████████████████▉     | 1054/1230 [7:58:36<1:18:56, 26.91s/it]
 86%|██████████████████████████████     | 1055/1230 [7:59:01<1:17:09, 26.46s/it]
                                                                                
{'loss': 0.667, 'grad_norm': 0.4044269323348999, 'learning_rate': 2.6530932420294822e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.89, 'epoch': 0.75}

 86%|██████████████████████████████     | 1055/1230 [7:59:01<1:17:09, 26.46s/it]
 86%|██████████████████████████████     | 1056/1230 [7:59:31<1:19:28, 27.40s/it]
                                                                                
{'loss': 0.7158, 'grad_norm': 0.4257482588291168, 'learning_rate': 2.6435979592817196e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.35, 'epoch': 0.75}

 86%|██████████████████████████████     | 1056/1230 [7:59:31<1:19:28, 27.40s/it]
 86%|██████████████████████████████     | 1057/1230 [8:00:01<1:20:57, 28.08s/it]
                                                                                
{'loss': 0.7163, 'grad_norm': 0.4057745635509491, 'learning_rate': 2.6341529122015456e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.92, 'epoch': 0.75}

 86%|██████████████████████████████     | 1057/1230 [8:00:01<1:20:57, 28.08s/it]
 86%|██████████████████████████████     | 1058/1230 [8:00:26<1:18:37, 27.42s/it]
                                                                                
{'loss': 0.7138, 'grad_norm': 0.4284358620643616, 'learning_rate': 2.6247581661764992e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.55, 'epoch': 0.75}

 86%|██████████████████████████████     | 1058/1230 [8:00:26<1:18:37, 27.42s/it]
 86%|██████████████████████████████▏    | 1059/1230 [8:00:53<1:17:06, 27.06s/it]
                                                                                
{'loss': 0.6579, 'grad_norm': 0.41423219442367554, 'learning_rate': 2.61541378624588e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.24, 'epoch': 0.75}

 86%|██████████████████████████████▏    | 1059/1230 [8:00:53<1:17:06, 27.06s/it]
 86%|██████████████████████████████▏    | 1060/1230 [8:01:18<1:14:56, 26.45s/it]
                                                                                
{'loss': 0.6758, 'grad_norm': 0.3957141935825348, 'learning_rate': 2.606119837100313e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.55, 'epoch': 0.75}

 86%|██████████████████████████████▏    | 1060/1230 [8:01:18<1:14:56, 26.45s/it]
 86%|██████████████████████████████▏    | 1061/1230 [8:01:43<1:13:26, 26.08s/it]
                                                                                
{'loss': 0.6983, 'grad_norm': 0.4324049651622772, 'learning_rate': 2.5968763830812893e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.85, 'epoch': 0.75}

 86%|██████████████████████████████▏    | 1061/1230 [8:01:43<1:13:26, 26.08s/it]
 86%|██████████████████████████████▏    | 1062/1230 [8:02:08<1:12:15, 25.81s/it]
                                                                                
{'loss': 0.6716, 'grad_norm': 0.39578551054000854, 'learning_rate': 2.5876834881807277e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.45, 'epoch': 0.75}

 86%|██████████████████████████████▏    | 1062/1230 [8:02:08<1:12:15, 25.81s/it]
 86%|██████████████████████████████▏    | 1063/1230 [8:02:33<1:11:17, 25.61s/it]
                                                                                
{'loss': 0.6467, 'grad_norm': 0.4307519197463989, 'learning_rate': 2.5785412160405304e-05, 'memory/max_active (GiB)': 9.25, 'memory/max_allocated (GiB)': 9.25, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.79, 'epoch': 0.75}

 86%|██████████████████████████████▏    | 1063/1230 [8:02:33<1:11:17, 25.61s/it]
 87%|██████████████████████████████▎    | 1064/1230 [8:02:59<1:10:38, 25.53s/it]
                                                                                
{'loss': 0.7037, 'grad_norm': 0.4196504056453705, 'learning_rate': 2.5694496299521352e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.94, 'epoch': 0.75}

 87%|██████████████████████████████▎    | 1064/1230 [8:02:59<1:10:38, 25.53s/it]
 87%|██████████████████████████████▎    | 1065/1230 [8:03:28<1:13:29, 26.72s/it]
                                                                                
{'loss': 0.6825, 'grad_norm': 0.41679373383522034, 'learning_rate': 2.560408792856088e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.06, 'epoch': 0.75}

 87%|██████████████████████████████▎    | 1065/1230 [8:03:28<1:13:29, 26.72s/it]
 87%|██████████████████████████████▎    | 1066/1230 [8:03:57<1:15:04, 27.47s/it]
                                                                                
{'loss': 0.6902, 'grad_norm': 0.41563695669174194, 'learning_rate': 2.5514187673416016e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.7, 'epoch': 0.75}

 87%|██████████████████████████████▎    | 1066/1230 [8:03:57<1:15:04, 27.47s/it]
 87%|██████████████████████████████▎    | 1067/1230 [8:04:27<1:16:19, 28.09s/it]
                                                                                
{'loss': 0.6655, 'grad_norm': 0.39381110668182373, 'learning_rate': 2.5424796156461212e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.47, 'epoch': 0.76}

 87%|██████████████████████████████▎    | 1067/1230 [8:04:27<1:16:19, 28.09s/it]
 87%|██████████████████████████████▍    | 1068/1230 [8:04:56<1:16:50, 28.46s/it]
                                                                                
{'loss': 0.723, 'grad_norm': 0.41844600439071655, 'learning_rate': 2.5335913996548985e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.91, 'epoch': 0.76}

 87%|██████████████████████████████▍    | 1068/1230 [8:04:56<1:16:50, 28.46s/it]
 87%|██████████████████████████████▍    | 1069/1230 [8:05:21<1:13:40, 27.46s/it]
                                                                                
{'loss': 0.6944, 'grad_norm': 0.39662569761276245, 'learning_rate': 2.5247541809005563e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.78, 'epoch': 0.76}

 87%|██████████████████████████████▍    | 1069/1230 [8:05:21<1:13:40, 27.46s/it]
 87%|██████████████████████████████▍    | 1070/1230 [8:05:50<1:14:28, 27.93s/it]
                                                                                
{'loss': 0.6641, 'grad_norm': 0.4135167896747589, 'learning_rate': 2.51596802056267e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.3, 'epoch': 0.76}

 87%|██████████████████████████████▍    | 1070/1230 [8:05:50<1:14:28, 27.93s/it]
 87%|██████████████████████████████▍    | 1071/1230 [8:06:15<1:11:50, 27.11s/it]
                                                                                
{'loss': 0.6878, 'grad_norm': 0.4124796390533447, 'learning_rate': 2.507232979467337e-05, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.45, 'epoch': 0.76}

 87%|██████████████████████████████▍    | 1071/1230 [8:06:15<1:11:50, 27.11s/it]
 87%|██████████████████████████████▌    | 1072/1230 [8:06:45<1:13:05, 27.75s/it]
                                                                                
{'loss': 0.6514, 'grad_norm': 0.3782827854156494, 'learning_rate': 2.4985491180867628e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.58, 'epoch': 0.76}

 87%|██████████████████████████████▌    | 1072/1230 [8:06:45<1:13:05, 27.75s/it]
 87%|██████████████████████████████▌    | 1073/1230 [8:07:14<1:13:43, 28.18s/it]
                                                                                
{'loss': 0.6688, 'grad_norm': 0.41098055243492126, 'learning_rate': 2.4899164965388355e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.66, 'epoch': 0.76}

 87%|██████████████████████████████▌    | 1073/1230 [8:07:14<1:13:43, 28.18s/it]
 87%|██████████████████████████████▌    | 1074/1230 [8:07:43<1:14:14, 28.55s/it]
                                                                                
{'loss': 0.6605, 'grad_norm': 0.3905984163284302, 'learning_rate': 2.481335174586713e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.63, 'epoch': 0.76}

 87%|██████████████████████████████▌    | 1074/1230 [8:07:43<1:14:14, 28.55s/it]
 87%|██████████████████████████████▌    | 1075/1230 [8:08:09<1:11:12, 27.57s/it]
                                                                                
{'loss': 0.7323, 'grad_norm': 0.4224952161312103, 'learning_rate': 2.4728052116384112e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.4, 'epoch': 0.76}

 87%|██████████████████████████████▌    | 1075/1230 [8:08:09<1:11:12, 27.57s/it]
 87%|██████████████████████████████▌    | 1076/1230 [8:08:38<1:12:13, 28.14s/it]
                                                                                
{'loss': 0.7296, 'grad_norm': 0.4247491955757141, 'learning_rate': 2.46432666674639e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.46, 'epoch': 0.76}

 87%|██████████████████████████████▌    | 1076/1230 [8:08:38<1:12:13, 28.14s/it]
 88%|██████████████████████████████▋    | 1077/1230 [8:09:03<1:09:39, 27.32s/it]
                                                                                
{'loss': 0.6863, 'grad_norm': 0.4035341441631317, 'learning_rate': 2.4558995986071392e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.34, 'epoch': 0.76}

 88%|██████████████████████████████▋    | 1077/1230 [8:09:03<1:09:39, 27.32s/it]
 88%|██████████████████████████████▋    | 1078/1230 [8:09:33<1:10:45, 27.93s/it]
                                                                                
{'loss': 0.6568, 'grad_norm': 0.410239040851593, 'learning_rate': 2.447524065560788e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.76, 'epoch': 0.76}

 88%|██████████████████████████████▋    | 1078/1230 [8:09:33<1:10:45, 27.93s/it]
 88%|██████████████████████████████▋    | 1079/1230 [8:09:59<1:08:44, 27.32s/it]
                                                                                
{'loss': 0.7032, 'grad_norm': 0.4438851773738861, 'learning_rate': 2.439200125590683e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.55, 'epoch': 0.76}

 88%|██████████████████████████████▋    | 1079/1230 [8:09:59<1:08:44, 27.32s/it]
 88%|██████████████████████████████▋    | 1080/1230 [8:10:28<1:09:54, 27.96s/it]
                                                                                
{'loss': 0.6725, 'grad_norm': 0.39562419056892395, 'learning_rate': 2.430927836322997e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.28, 'epoch': 0.76}

 88%|██████████████████████████████▋    | 1080/1230 [8:10:28<1:09:54, 27.96s/it]
 88%|██████████████████████████████▊    | 1081/1230 [8:10:57<1:10:15, 28.29s/it]
                                                                                
{'loss': 0.6859, 'grad_norm': 0.39278972148895264, 'learning_rate': 2.4227072550263293e-05, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 147.3, 'epoch': 0.77}

 88%|██████████████████████████████▊    | 1081/1230 [8:10:57<1:10:15, 28.29s/it]
 88%|██████████████████████████████▊    | 1082/1230 [8:11:26<1:10:23, 28.54s/it]
                                                                                
{'loss': 0.6596, 'grad_norm': 0.40861329436302185, 'learning_rate': 2.414538438611306e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.83, 'epoch': 0.77}

 88%|██████████████████████████████▊    | 1082/1230 [8:11:26<1:10:23, 28.54s/it]
 88%|██████████████████████████████▊    | 1083/1230 [8:11:52<1:07:47, 27.67s/it]
                                                                                
{'loss': 0.6977, 'grad_norm': 0.41427889466285706, 'learning_rate': 2.4064214436301905e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 199.88, 'epoch': 0.77}

 88%|██████████████████████████████▊    | 1083/1230 [8:11:52<1:07:47, 27.67s/it]
 88%|██████████████████████████████▊    | 1084/1230 [8:12:18<1:06:00, 27.12s/it]
                                                                                
{'loss': 0.729, 'grad_norm': 0.4234538674354553, 'learning_rate': 2.3983563262764874e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.11, 'epoch': 0.77}

 88%|██████████████████████████████▊    | 1084/1230 [8:12:18<1:06:00, 27.12s/it]
 88%|██████████████████████████████▊    | 1085/1230 [8:12:47<1:07:12, 27.81s/it]
                                                                                
{'loss': 0.6909, 'grad_norm': 0.40957313776016235, 'learning_rate': 2.390343142384553e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.41, 'epoch': 0.77}

 88%|██████████████████████████████▊    | 1085/1230 [8:12:47<1:07:12, 27.81s/it]
 88%|██████████████████████████████▉    | 1086/1230 [8:13:17<1:08:06, 28.38s/it]
                                                                                
{'loss': 0.6932, 'grad_norm': 0.4153284728527069, 'learning_rate': 2.382381947429215e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.25, 'epoch': 0.77}

 88%|██████████████████████████████▉    | 1086/1230 [8:13:17<1:08:06, 28.38s/it]
 88%|██████████████████████████████▉    | 1087/1230 [8:13:42<1:05:20, 27.41s/it]
                                                                                
{'loss': 0.6695, 'grad_norm': 0.43017083406448364, 'learning_rate': 2.3744727965253824e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.83, 'epoch': 0.77}

 88%|██████████████████████████████▉    | 1087/1230 [8:13:42<1:05:20, 27.41s/it]
 88%|██████████████████████████████▉    | 1088/1230 [8:14:07<1:03:11, 26.70s/it]
                                                                                
{'loss': 0.6908, 'grad_norm': 0.40768104791641235, 'learning_rate': 2.3666157444276688e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.79, 'epoch': 0.77}

 88%|██████████████████████████████▉    | 1088/1230 [8:14:07<1:03:11, 26.70s/it]
 89%|██████████████████████████████▉    | 1089/1230 [8:14:32<1:01:32, 26.19s/it]
                                                                                
{'loss': 0.7193, 'grad_norm': 0.4171963036060333, 'learning_rate': 2.358810845530007e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.51, 'epoch': 0.77}

 89%|██████████████████████████████▉    | 1089/1230 [8:14:32<1:01:32, 26.19s/it]
 89%|███████████████████████████████    | 1090/1230 [8:15:01<1:03:03, 27.03s/it]
                                                                                
{'loss': 0.7305, 'grad_norm': 0.4062877297401428, 'learning_rate': 2.3510581538652788e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.1, 'epoch': 0.77}

 89%|███████████████████████████████    | 1090/1230 [8:15:01<1:03:03, 27.03s/it]
 89%|███████████████████████████████    | 1091/1230 [8:15:26<1:01:16, 26.45s/it]
                                                                                
{'loss': 0.7018, 'grad_norm': 0.41952499747276306, 'learning_rate': 2.3433577231049354e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 189.23, 'epoch': 0.77}

 89%|███████████████████████████████    | 1091/1230 [8:15:26<1:01:16, 26.45s/it]
 89%|███████████████████████████████    | 1092/1230 [8:15:56<1:02:52, 27.34s/it]
                                                                                
{'loss': 0.6616, 'grad_norm': 0.40680932998657227, 'learning_rate': 2.335709606558632e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 144.41, 'epoch': 0.77}

 89%|███████████████████████████████    | 1092/1230 [8:15:56<1:02:52, 27.34s/it]
 89%|███████████████████████████████    | 1093/1230 [8:16:21<1:01:18, 26.85s/it]
                                                                                
{'loss': 0.6953, 'grad_norm': 0.4129028022289276, 'learning_rate': 2.3281138571738536e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.17, 'epoch': 0.77}

 89%|███████████████████████████████    | 1093/1230 [8:16:21<1:01:18, 26.85s/it]
 89%|███████████████████████████████▏   | 1094/1230 [8:16:51<1:02:43, 27.67s/it]
                                                                                
{'loss': 0.6764, 'grad_norm': 0.3931202292442322, 'learning_rate': 2.3205705275355514e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 151.68, 'epoch': 0.77}

 89%|███████████████████████████████▏   | 1094/1230 [8:16:51<1:02:43, 27.67s/it]
 89%|███████████████████████████████▏   | 1095/1230 [8:17:16<1:00:39, 26.96s/it]
                                                                                
{'loss': 0.6927, 'grad_norm': 0.4297466278076172, 'learning_rate': 2.3130796698657757e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.85, 'epoch': 0.77}

 89%|███████████████████████████████▏   | 1095/1230 [8:17:16<1:00:39, 26.96s/it]
 89%|███████████████████████████████▏   | 1096/1230 [8:17:46<1:01:53, 27.71s/it]
                                                                                
{'loss': 0.6675, 'grad_norm': 0.38729625940322876, 'learning_rate': 2.3056413360233176e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.8, 'epoch': 0.78}

 89%|███████████████████████████████▏   | 1096/1230 [8:17:46<1:01:53, 27.71s/it]
 89%|████████████████████████████████▉    | 1097/1230 [8:18:11<59:58, 27.06s/it]
                                                                                
{'loss': 0.7133, 'grad_norm': 0.4064250588417053, 'learning_rate': 2.2982555775033483e-05, 'memory/max_active (GiB)': 9.62, 'memory/max_allocated (GiB)': 9.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 209.15, 'epoch': 0.78}

 89%|████████████████████████████████▉    | 1097/1230 [8:18:11<59:58, 27.06s/it]
 89%|█████████████████████████████████    | 1098/1230 [8:18:37<58:23, 26.54s/it]
                                                                                
{'loss': 0.6765, 'grad_norm': 0.40771999955177307, 'learning_rate': 2.2909224454370628e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.56, 'epoch': 0.78}

 89%|█████████████████████████████████    | 1098/1230 [8:18:37<58:23, 26.54s/it]
 89%|█████████████████████████████████    | 1099/1230 [8:19:06<59:59, 27.48s/it]
                                                                                
{'loss': 0.6438, 'grad_norm': 0.39461106061935425, 'learning_rate': 2.283641990591329e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.61, 'epoch': 0.78}

 89%|█████████████████████████████████    | 1099/1230 [8:19:06<59:59, 27.48s/it]
 89%|███████████████████████████████▎   | 1100/1230 [8:19:36<1:00:56, 28.13s/it]
                                                                                
{'loss': 0.7113, 'grad_norm': 0.4447220265865326, 'learning_rate': 2.2764142633683267e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.04, 'epoch': 0.78}

 89%|███████████████████████████████▎   | 1100/1230 [8:19:36<1:00:56, 28.13s/it]
 90%|█████████████████████████████████    | 1101/1230 [8:20:02<58:53, 27.39s/it]
                                                                                
{'loss': 0.6902, 'grad_norm': 0.41594141721725464, 'learning_rate': 2.2692393138052107e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.74, 'epoch': 0.78}

 90%|█████████████████████████████████    | 1101/1230 [8:20:02<58:53, 27.39s/it]
 90%|█████████████████████████████████▏   | 1102/1230 [8:20:27<57:02, 26.74s/it]
                                                                                
{'loss': 0.6462, 'grad_norm': 0.39971351623535156, 'learning_rate': 2.26211719157376e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.74, 'epoch': 0.78}

 90%|█████████████████████████████████▏   | 1102/1230 [8:20:27<57:02, 26.74s/it]
 90%|█████████████████████████████████▏   | 1103/1230 [8:20:53<56:20, 26.62s/it]
                                                                                
{'loss': 0.6625, 'grad_norm': 0.38768643140792847, 'learning_rate': 2.255047945980028e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.6, 'epoch': 0.78}

 90%|█████████████████████████████████▏   | 1103/1230 [8:20:53<56:20, 26.62s/it]
 90%|█████████████████████████████████▏   | 1104/1230 [8:21:19<55:08, 26.26s/it]
                                                                                
{'loss': 0.6697, 'grad_norm': 0.41423314809799194, 'learning_rate': 2.2480316259640097e-05, 'memory/max_active (GiB)': 12.05, 'memory/max_allocated (GiB)': 12.05, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.02, 'epoch': 0.78}

 90%|█████████████████████████████████▏   | 1104/1230 [8:21:19<55:08, 26.26s/it]
 90%|█████████████████████████████████▏   | 1105/1230 [8:21:45<54:40, 26.25s/it]
                                                                                
{'loss': 0.6841, 'grad_norm': 0.43325287103652954, 'learning_rate': 2.2410682800992976e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.85, 'epoch': 0.78}

 90%|█████████████████████████████████▏   | 1105/1230 [8:21:45<54:40, 26.25s/it]
 90%|█████████████████████████████████▎   | 1106/1230 [8:22:14<56:11, 27.19s/it]
                                                                                
{'loss': 0.6778, 'grad_norm': 0.4000975787639618, 'learning_rate': 2.2341579565927475e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.6, 'epoch': 0.78}

 90%|█████████████████████████████████▎   | 1106/1230 [8:22:14<56:11, 27.19s/it]
 90%|█████████████████████████████████▎   | 1107/1230 [8:22:44<57:05, 27.85s/it]
                                                                                
{'loss': 0.6949, 'grad_norm': 0.41182824969291687, 'learning_rate': 2.2273007032841447e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 161.7, 'epoch': 0.78}

 90%|█████████████████████████████████▎   | 1107/1230 [8:22:44<57:05, 27.85s/it]
 90%|█████████████████████████████████▎   | 1108/1230 [8:23:09<55:04, 27.08s/it]
                                                                                
{'loss': 0.7004, 'grad_norm': 0.41061943769454956, 'learning_rate': 2.220496567645873e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.2, 'epoch': 0.78}

 90%|█████████████████████████████████▎   | 1108/1230 [8:23:09<55:04, 27.08s/it]
 90%|█████████████████████████████████▎   | 1109/1230 [8:23:34<53:28, 26.52s/it]
                                                                                
{'loss': 0.6853, 'grad_norm': 0.4184601604938507, 'learning_rate': 2.213745596782586e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.64, 'epoch': 0.78}

 90%|█████████████████████████████████▎   | 1109/1230 [8:23:34<53:28, 26.52s/it]
 90%|█████████████████████████████████▍   | 1110/1230 [8:24:00<52:53, 26.44s/it]
                                                                                
{'loss': 0.6622, 'grad_norm': 0.39949801564216614, 'learning_rate': 2.2070478374308757e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.89, 'epoch': 0.79}

 90%|█████████████████████████████████▍   | 1110/1230 [8:24:00<52:53, 26.44s/it]
 90%|█████████████████████████████████▍   | 1111/1230 [8:24:26<51:50, 26.14s/it]
                                                                                
{'loss': 0.6687, 'grad_norm': 0.4185310900211334, 'learning_rate': 2.2004033359589594e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.38, 'epoch': 0.79}

 90%|█████████████████████████████████▍   | 1111/1230 [8:24:26<51:50, 26.14s/it]
 90%|█████████████████████████████████▍   | 1112/1230 [8:24:51<50:54, 25.88s/it]
                                                                                
{'loss': 0.6429, 'grad_norm': 0.41338810324668884, 'learning_rate': 2.1938121383663516e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 187.09, 'epoch': 0.79}

 90%|█████████████████████████████████▍   | 1112/1230 [8:24:51<50:54, 25.88s/it]
 90%|█████████████████████████████████▍   | 1113/1230 [8:25:20<52:34, 26.96s/it]
                                                                                
{'loss': 0.6729, 'grad_norm': 0.42288926243782043, 'learning_rate': 2.187274290283546e-05, 'memory/max_active (GiB)': 10.78, 'memory/max_allocated (GiB)': 10.78, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 147.13, 'epoch': 0.79}

 90%|█████████████████████████████████▍   | 1113/1230 [8:25:20<52:34, 26.96s/it]
 91%|█████████████████████████████████▌   | 1114/1230 [8:25:50<53:31, 27.69s/it]
                                                                                
{'loss': 0.6945, 'grad_norm': 0.4068908989429474, 'learning_rate': 2.1807898369717026e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.51, 'epoch': 0.79}

 91%|█████████████████████████████████▌   | 1114/1230 [8:25:50<53:31, 27.69s/it]
 91%|█████████████████████████████████▌   | 1115/1230 [8:26:19<54:00, 28.18s/it]
                                                                                
{'loss': 0.7123, 'grad_norm': 0.4213281273841858, 'learning_rate': 2.1743588233223295e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.26, 'epoch': 0.79}

 91%|█████████████████████████████████▌   | 1115/1230 [8:26:19<54:00, 28.18s/it]
 91%|█████████████████████████████████▌   | 1116/1230 [8:26:45<52:20, 27.55s/it]
                                                                                
{'loss': 0.7113, 'grad_norm': 0.43575406074523926, 'learning_rate': 2.167981293856975e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.27, 'epoch': 0.79}

 91%|█████████████████████████████████▌   | 1116/1230 [8:26:45<52:20, 27.55s/it]
 91%|█████████████████████████████████▌   | 1117/1230 [8:27:10<50:23, 26.76s/it]
                                                                                
{'loss': 0.6669, 'grad_norm': 0.4063262939453125, 'learning_rate': 2.1616572927269223e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.29, 'epoch': 0.79}

 91%|█████████████████████████████████▌   | 1117/1230 [8:27:10<50:23, 26.76s/it]
 91%|█████████████████████████████████▋   | 1118/1230 [8:27:40<51:31, 27.60s/it]
                                                                                
{'loss': 0.6872, 'grad_norm': 0.4225292503833771, 'learning_rate': 2.1553868637128794e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 160.54, 'epoch': 0.79}

 91%|█████████████████████████████████▋   | 1118/1230 [8:27:40<51:31, 27.60s/it]
 91%|█████████████████████████████████▋   | 1119/1230 [8:28:06<50:06, 27.09s/it]
                                                                                
{'loss': 0.6765, 'grad_norm': 0.3985667824745178, 'learning_rate': 2.1491700502246786e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.55, 'epoch': 0.79}

 91%|█████████████████████████████████▋   | 1119/1230 [8:28:06<50:06, 27.09s/it]
 91%|█████████████████████████████████▋   | 1120/1230 [8:28:31<48:27, 26.43s/it]
                                                                                
{'loss': 0.653, 'grad_norm': 0.41316065192222595, 'learning_rate': 2.1430068953009714e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.37, 'epoch': 0.79}

 91%|█████████████████████████████████▋   | 1120/1230 [8:28:31<48:27, 26.43s/it]
 91%|█████████████████████████████████▋   | 1121/1230 [8:28:55<47:12, 25.98s/it]
                                                                                
{'loss': 0.7109, 'grad_norm': 0.42526572942733765, 'learning_rate': 2.136897441608937e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.54, 'epoch': 0.79}

 91%|█████████████████████████████████▋   | 1121/1230 [8:28:55<47:12, 25.98s/it]
 91%|█████████████████████████████████▊   | 1122/1230 [8:29:21<46:17, 25.71s/it]
                                                                                
{'loss': 0.6684, 'grad_norm': 0.42239633202552795, 'learning_rate': 2.1308417314439847e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.11, 'epoch': 0.79}

 91%|█████████████████████████████████▊   | 1122/1230 [8:29:21<46:17, 25.71s/it]
 91%|█████████████████████████████████▊   | 1123/1230 [8:29:45<45:26, 25.48s/it]
                                                                                
{'loss': 0.6889, 'grad_norm': 0.4210778772830963, 'learning_rate': 2.1248398067294578e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.52, 'epoch': 0.79}

 91%|█████████████████████████████████▊   | 1123/1230 [8:29:46<45:26, 25.48s/it]
 91%|█████████████████████████████████▊   | 1124/1230 [8:30:11<44:49, 25.37s/it]
                                                                                
{'loss': 0.7107, 'grad_norm': 0.42994481325149536, 'learning_rate': 2.1188917090163496e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.79, 'epoch': 0.8}

 91%|█████████████████████████████████▊   | 1124/1230 [8:30:11<44:49, 25.37s/it]
 91%|█████████████████████████████████▊   | 1125/1230 [8:30:36<44:36, 25.49s/it]
                                                                                
{'loss': 0.6495, 'grad_norm': 0.40192756056785583, 'learning_rate': 2.112997479483005e-05, 'memory/max_active (GiB)': 11.15, 'memory/max_allocated (GiB)': 11.15, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.41, 'epoch': 0.8}

 91%|█████████████████████████████████▊   | 1125/1230 [8:30:36<44:36, 25.49s/it]
 92%|█████████████████████████████████▊   | 1126/1230 [8:31:01<43:55, 25.35s/it]
                                                                                
{'loss': 0.6897, 'grad_norm': 0.44515150785446167, 'learning_rate': 2.1071571589348514e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.4, 'epoch': 0.8}

 92%|█████████████████████████████████▊   | 1126/1230 [8:31:01<43:55, 25.35s/it]
 92%|█████████████████████████████████▉   | 1127/1230 [8:31:31<45:39, 26.59s/it]
                                                                                
{'loss': 0.674, 'grad_norm': 0.4171006679534912, 'learning_rate': 2.1013707878041e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 150.82, 'epoch': 0.8}

 92%|█████████████████████████████████▉   | 1127/1230 [8:31:31<45:39, 26.59s/it]
 92%|█████████████████████████████████▉   | 1128/1230 [8:32:00<46:33, 27.39s/it]
                                                                                
{'loss': 0.6944, 'grad_norm': 0.4009528160095215, 'learning_rate': 2.0956384061494776e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.47, 'epoch': 0.8}

 92%|█████████████████████████████████▉   | 1128/1230 [8:32:00<46:33, 27.39s/it]
 92%|█████████████████████████████████▉   | 1129/1230 [8:32:26<45:08, 26.82s/it]
                                                                                
{'loss': 0.7226, 'grad_norm': 0.43257954716682434, 'learning_rate': 2.0899600536559457e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.7, 'epoch': 0.8}

 92%|█████████████████████████████████▉   | 1129/1230 [8:32:26<45:08, 26.82s/it]
 92%|█████████████████████████████████▉   | 1130/1230 [8:32:51<43:57, 26.37s/it]
                                                                                
{'loss': 0.6843, 'grad_norm': 0.4471687078475952, 'learning_rate': 2.0843357696344187e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.58, 'epoch': 0.8}

 92%|█████████████████████████████████▉   | 1130/1230 [8:32:51<43:57, 26.37s/it]
 92%|██████████████████████████████████   | 1131/1230 [8:33:17<43:21, 26.28s/it]
                                                                                
{'loss': 0.6742, 'grad_norm': 0.3997613191604614, 'learning_rate': 2.0787655930215033e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.74, 'epoch': 0.8}

 92%|██████████████████████████████████   | 1131/1230 [8:33:17<43:21, 26.28s/it]
 92%|██████████████████████████████████   | 1132/1230 [8:33:42<42:26, 25.98s/it]
                                                                                
{'loss': 0.6448, 'grad_norm': 0.40072178840637207, 'learning_rate': 2.0732495623792235e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.89, 'epoch': 0.8}

 92%|██████████████████████████████████   | 1132/1230 [8:33:42<42:26, 25.98s/it]
 92%|██████████████████████████████████   | 1133/1230 [8:34:08<42:05, 26.04s/it]
                                                                                
{'loss': 0.6651, 'grad_norm': 0.39769604802131653, 'learning_rate': 2.067787715894751e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.58, 'epoch': 0.8}

 92%|██████████████████████████████████   | 1133/1230 [8:34:08<42:05, 26.04s/it]
 92%|██████████████████████████████████   | 1134/1230 [8:34:38<43:07, 26.96s/it]
                                                                                
{'loss': 0.7427, 'grad_norm': 0.39213672280311584, 'learning_rate': 2.062380091380147e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 169.9, 'epoch': 0.8}

 92%|██████████████████████████████████   | 1134/1230 [8:34:38<43:07, 26.96s/it]
 92%|██████████████████████████████████▏  | 1135/1230 [8:35:07<43:48, 27.67s/it]
                                                                                
{'loss': 0.6847, 'grad_norm': 0.41411083936691284, 'learning_rate': 2.057026726272092e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.35, 'epoch': 0.8}

 92%|██████████████████████████████████▏  | 1135/1230 [8:35:07<43:48, 27.67s/it]
 92%|██████████████████████████████████▏  | 1136/1230 [8:35:36<44:01, 28.10s/it]
                                                                                
{'loss': 0.715, 'grad_norm': 0.38596290349960327, 'learning_rate': 2.051727657631638e-05, 'memory/max_active (GiB)': 9.24, 'memory/max_allocated (GiB)': 9.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.48, 'epoch': 0.8}

 92%|██████████████████████████████████▏  | 1136/1230 [8:35:36<44:01, 28.10s/it]
 92%|██████████████████████████████████▏  | 1137/1230 [8:36:01<42:16, 27.27s/it]
                                                                                
{'loss': 0.6893, 'grad_norm': 0.42306122183799744, 'learning_rate': 2.0464829221439404e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.28, 'epoch': 0.8}

 92%|██████████████████████████████████▏  | 1137/1230 [8:36:01<42:16, 27.27s/it]
 93%|██████████████████████████████████▏  | 1138/1230 [8:36:27<41:00, 26.74s/it]
                                                                                
{'loss': 0.6729, 'grad_norm': 0.4252074956893921, 'learning_rate': 2.041292556118012e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.97, 'epoch': 0.81}

 93%|██████████████████████████████████▏  | 1138/1230 [8:36:27<41:00, 26.74s/it]
 93%|██████████████████████████████████▎  | 1139/1230 [8:36:56<41:45, 27.53s/it]
                                                                                
{'loss': 0.7021, 'grad_norm': 0.4235657751560211, 'learning_rate': 2.036156595486468e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.75, 'epoch': 0.81}

 93%|██████████████████████████████████▎  | 1139/1230 [8:36:56<41:45, 27.53s/it]
 93%|██████████████████████████████████▎  | 1140/1230 [8:37:26<42:04, 28.05s/it]
                                                                                
{'loss': 0.6882, 'grad_norm': 0.3945672810077667, 'learning_rate': 2.0310750758052775e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.63, 'epoch': 0.81}

 93%|██████████████████████████████████▎  | 1140/1230 [8:37:26<42:04, 28.05s/it]
 93%|██████████████████████████████████▎  | 1141/1230 [8:37:51<40:15, 27.14s/it]
                                                                                
{'loss': 0.6834, 'grad_norm': 0.3975962698459625, 'learning_rate': 2.0260480322535187e-05, 'memory/max_active (GiB)': 9.39, 'memory/max_allocated (GiB)': 9.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.51, 'epoch': 0.81}

 93%|██████████████████████████████████▎  | 1141/1230 [8:37:51<40:15, 27.14s/it]
 93%|██████████████████████████████████▎  | 1142/1230 [8:38:17<39:19, 26.82s/it]
                                                                                
{'loss': 0.6649, 'grad_norm': 0.420935720205307, 'learning_rate': 2.0210754996331355e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.42, 'epoch': 0.81}

 93%|██████████████████████████████████▎  | 1142/1230 [8:38:17<39:19, 26.82s/it]
 93%|██████████████████████████████████▍  | 1143/1230 [8:38:42<38:28, 26.53s/it]
                                                                                
{'loss': 0.6727, 'grad_norm': 0.3890937864780426, 'learning_rate': 2.0161575123686924e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.83, 'epoch': 0.81}

 93%|██████████████████████████████████▍  | 1143/1230 [8:38:42<38:28, 26.53s/it]
 93%|██████████████████████████████████▍  | 1144/1230 [8:39:12<39:11, 27.35s/it]
                                                                                
{'loss': 0.7276, 'grad_norm': 0.4327520430088043, 'learning_rate': 2.011294104507145e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.66, 'epoch': 0.81}

 93%|██████████████████████████████████▍  | 1144/1230 [8:39:12<39:11, 27.35s/it]
 93%|██████████████████████████████████▍  | 1145/1230 [8:39:37<37:47, 26.68s/it]
                                                                                
{'loss': 0.6897, 'grad_norm': 0.41321831941604614, 'learning_rate': 2.0064853097175916e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.47, 'epoch': 0.81}

 93%|██████████████████████████████████▍  | 1145/1230 [8:39:37<37:47, 26.68s/it]
 93%|██████████████████████████████████▍  | 1146/1230 [8:40:03<37:03, 26.47s/it]
                                                                                
{'loss': 0.6947, 'grad_norm': 0.4077063202857971, 'learning_rate': 2.0017311612910526e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.62, 'epoch': 0.81}

 93%|██████████████████████████████████▍  | 1146/1230 [8:40:03<37:03, 26.47s/it]
 93%|██████████████████████████████████▌  | 1147/1230 [8:40:28<36:03, 26.07s/it]
                                                                                
{'loss': 0.7017, 'grad_norm': 0.4294396936893463, 'learning_rate': 1.997031692140234e-05, 'memory/max_active (GiB)': 10.02, 'memory/max_allocated (GiB)': 10.02, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.07, 'epoch': 0.81}

 93%|██████████████████████████████████▌  | 1147/1230 [8:40:28<36:03, 26.07s/it]
 93%|██████████████████████████████████▌  | 1148/1230 [8:40:57<36:54, 27.01s/it]
                                                                                
{'loss': 0.6643, 'grad_norm': 0.4129369258880615, 'learning_rate': 1.9923869347992972e-05, 'memory/max_active (GiB)': 10.53, 'memory/max_allocated (GiB)': 10.53, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.18, 'epoch': 0.81}

 93%|██████████████████████████████████▌  | 1148/1230 [8:40:57<36:54, 27.01s/it]
 93%|██████████████████████████████████▌  | 1149/1230 [8:41:23<35:50, 26.55s/it]
                                                                                
{'loss': 0.705, 'grad_norm': 0.42129021883010864, 'learning_rate': 1.9877969214236422e-05, 'memory/max_active (GiB)': 9.77, 'memory/max_allocated (GiB)': 9.77, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.19, 'epoch': 0.81}

 93%|██████████████████████████████████▌  | 1149/1230 [8:41:23<35:50, 26.55s/it]
 93%|██████████████████████████████████▌  | 1150/1230 [8:41:49<35:10, 26.38s/it]
                                                                                
{'loss': 0.6401, 'grad_norm': 0.3770723044872284, 'learning_rate': 1.983261683789672e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.94, 'epoch': 0.81}

 93%|██████████████████████████████████▌  | 1150/1230 [8:41:49<35:10, 26.38s/it]
 94%|██████████████████████████████████▌  | 1151/1230 [8:42:14<34:19, 26.07s/it]
                                                                                
{'loss': 0.6913, 'grad_norm': 0.4425869286060333, 'learning_rate': 1.9787812532945863e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.7, 'epoch': 0.81}

 94%|██████████████████████████████████▌  | 1151/1230 [8:42:14<34:19, 26.07s/it]
 94%|██████████████████████████████████▋  | 1152/1230 [8:42:39<33:36, 25.85s/it]
                                                                                
{'loss': 0.6472, 'grad_norm': 0.44125741720199585, 'learning_rate': 1.9743556609561567e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 185.6, 'epoch': 0.82}

 94%|██████████████████████████████████▋  | 1152/1230 [8:42:39<33:36, 25.85s/it]
 94%|██████████████████████████████████▋  | 1153/1230 [8:43:09<34:30, 26.89s/it]
                                                                                
{'loss': 0.7004, 'grad_norm': 0.4102002680301666, 'learning_rate': 1.9699849374125127e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.39, 'epoch': 0.82}

 94%|██████████████████████████████████▋  | 1153/1230 [8:43:09<34:30, 26.89s/it]
 94%|██████████████████████████████████▋  | 1154/1230 [8:43:38<34:49, 27.49s/it]
                                                                                
{'loss': 0.6887, 'grad_norm': 0.41456976532936096, 'learning_rate': 1.9656691129219294e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.32, 'epoch': 0.82}

 94%|██████████████████████████████████▋  | 1154/1230 [8:43:38<34:49, 27.49s/it]
 94%|██████████████████████████████████▋  | 1155/1230 [8:44:02<33:24, 26.73s/it]
                                                                                
{'loss': 0.6777, 'grad_norm': 0.43072882294654846, 'learning_rate': 1.96140821736262e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.12, 'epoch': 0.82}

 94%|██████████████████████████████████▋  | 1155/1230 [8:44:02<33:24, 26.73s/it]
 94%|██████████████████████████████████▊  | 1156/1230 [8:44:31<33:46, 27.38s/it]
                                                                                
{'loss': 0.6825, 'grad_norm': 0.4004732072353363, 'learning_rate': 1.9572022802325273e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 174.13, 'epoch': 0.82}

 94%|██████████████████████████████████▊  | 1156/1230 [8:44:31<33:46, 27.38s/it]
 94%|██████████████████████████████████▊  | 1157/1230 [8:44:56<32:24, 26.64s/it]
                                                                                
{'loss': 0.6359, 'grad_norm': 0.3940545916557312, 'learning_rate': 1.9530513306491192e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.64, 'epoch': 0.82}

 94%|██████████████████████████████████▊  | 1157/1230 [8:44:56<32:24, 26.64s/it]
 94%|██████████████████████████████████▊  | 1158/1230 [8:45:26<32:57, 27.47s/it]
                                                                                
{'loss': 0.6868, 'grad_norm': 0.39971596002578735, 'learning_rate': 1.9489553973491912e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 164.89, 'epoch': 0.82}

 94%|██████████████████████████████████▊  | 1158/1230 [8:45:26<32:57, 27.47s/it]
 94%|██████████████████████████████████▊  | 1159/1230 [8:45:52<31:57, 27.01s/it]
                                                                                
{'loss': 0.6805, 'grad_norm': 0.40119826793670654, 'learning_rate': 1.944914508688659e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 192.17, 'epoch': 0.82}

 94%|██████████████████████████████████▊  | 1159/1230 [8:45:52<31:57, 27.01s/it]
 94%|██████████████████████████████████▉  | 1160/1230 [8:46:21<32:22, 27.75s/it]
                                                                                
{'loss': 0.655, 'grad_norm': 0.38894787430763245, 'learning_rate': 1.940928692642371e-05, 'memory/max_active (GiB)': 9.37, 'memory/max_allocated (GiB)': 9.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.42, 'epoch': 0.82}

 94%|██████████████████████████████████▉  | 1160/1230 [8:46:21<32:22, 27.75s/it]
 94%|██████████████████████████████████▉  | 1161/1230 [8:46:47<31:08, 27.08s/it]
                                                                                
{'loss': 0.6487, 'grad_norm': 0.43726229667663574, 'learning_rate': 1.936997976803911e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.44, 'epoch': 0.82}

 94%|██████████████████████████████████▉  | 1161/1230 [8:46:47<31:08, 27.08s/it]
 94%|██████████████████████████████████▉  | 1162/1230 [8:47:16<31:34, 27.85s/it]
                                                                                
{'loss': 0.6951, 'grad_norm': 0.42368245124816895, 'learning_rate': 1.9331223883854044e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.2, 'epoch': 0.82}

 94%|██████████████████████████████████▉  | 1162/1230 [8:47:16<31:34, 27.85s/it]
 95%|██████████████████████████████████▉  | 1163/1230 [8:47:42<30:17, 27.12s/it]
                                                                                
{'loss': 0.7045, 'grad_norm': 0.41609612107276917, 'learning_rate': 1.9293019542173343e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.75, 'epoch': 0.82}

 95%|██████████████████████████████████▉  | 1163/1230 [8:47:42<30:17, 27.12s/it]
 95%|███████████████████████████████████  | 1164/1230 [8:48:08<29:33, 26.87s/it]
                                                                                
{'loss': 0.6957, 'grad_norm': 0.41759195923805237, 'learning_rate': 1.9255367007483555e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.75, 'epoch': 0.82}

 95%|███████████████████████████████████  | 1164/1230 [8:48:08<29:33, 26.87s/it]
 95%|███████████████████████████████████  | 1165/1230 [8:48:37<29:54, 27.61s/it]
                                                                                
{'loss': 0.6678, 'grad_norm': 0.4016987085342407, 'learning_rate': 1.921826654045106e-05, 'memory/max_active (GiB)': 10.89, 'memory/max_allocated (GiB)': 10.89, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.75, 'epoch': 0.82}

 95%|███████████████████████████████████  | 1165/1230 [8:48:37<29:54, 27.61s/it]
 95%|███████████████████████████████████  | 1166/1230 [8:49:02<28:34, 26.79s/it]
                                                                                
{'loss': 0.6614, 'grad_norm': 0.4346320927143097, 'learning_rate': 1.9181718397920336e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.29, 'epoch': 0.83}

 95%|███████████████████████████████████  | 1166/1230 [8:49:02<28:34, 26.79s/it]
 95%|███████████████████████████████████  | 1167/1230 [8:49:31<28:45, 27.39s/it]
                                                                                
{'loss': 0.6875, 'grad_norm': 0.3928740322589874, 'learning_rate': 1.914572283291215e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.44, 'epoch': 0.83}

 95%|███████████████████████████████████  | 1167/1230 [8:49:31<28:45, 27.39s/it]
 95%|███████████████████████████████████▏ | 1168/1230 [8:49:57<27:47, 26.90s/it]
                                                                                
{'loss': 0.6735, 'grad_norm': 0.422498881816864, 'learning_rate': 1.911028009462178e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 191.93, 'epoch': 0.83}

 95%|███████████████████████████████████▏ | 1168/1230 [8:49:57<27:47, 26.90s/it]
 95%|███████████████████████████████████▏ | 1169/1230 [8:50:26<28:02, 27.59s/it]
                                                                                
{'loss': 0.7041, 'grad_norm': 0.4218270778656006, 'learning_rate': 1.9075390428417352e-05, 'memory/max_active (GiB)': 8.24, 'memory/max_allocated (GiB)': 8.24, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.29, 'epoch': 0.83}

 95%|███████████████████████████████████▏ | 1169/1230 [8:50:26<28:02, 27.59s/it]
 95%|███████████████████████████████████▏ | 1170/1230 [8:50:55<28:10, 28.17s/it]
                                                                                
{'loss': 0.6895, 'grad_norm': 0.43301835656166077, 'learning_rate': 1.904105407583807e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.15, 'epoch': 0.83}

 95%|███████████████████████████████████▏ | 1170/1230 [8:50:55<28:10, 28.17s/it]
 95%|███████████████████████████████████▏ | 1171/1230 [8:51:21<26:52, 27.32s/it]
                                                                                
{'loss': 0.7272, 'grad_norm': 0.4519399106502533, 'learning_rate': 1.9007271274592578e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 188.9, 'epoch': 0.83}

 95%|███████████████████████████████████▏ | 1171/1230 [8:51:21<26:52, 27.32s/it]
 95%|███████████████████████████████████▎ | 1172/1230 [8:51:46<25:49, 26.72s/it]
                                                                                
{'loss': 0.6613, 'grad_norm': 0.42387619614601135, 'learning_rate': 1.8974042258557363e-05, 'memory/max_active (GiB)': 8.75, 'memory/max_allocated (GiB)': 8.75, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.88, 'epoch': 0.83}

 95%|███████████████████████████████████▎ | 1172/1230 [8:51:46<25:49, 26.72s/it]
 95%|███████████████████████████████████▎ | 1173/1230 [8:52:12<25:07, 26.45s/it]
                                                                                
{'loss': 0.7042, 'grad_norm': 0.418804794549942, 'learning_rate': 1.8941367257775027e-05, 'memory/max_active (GiB)': 10.39, 'memory/max_allocated (GiB)': 10.39, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 197.21, 'epoch': 0.83}

 95%|███████████████████████████████████▎ | 1173/1230 [8:52:12<25:07, 26.45s/it]
 95%|███████████████████████████████████▎ | 1174/1230 [8:52:42<25:38, 27.47s/it]
                                                                                
{'loss': 0.6701, 'grad_norm': 0.40059778094291687, 'learning_rate': 1.8909246498452802e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.08, 'epoch': 0.83}

 95%|███████████████████████████████████▎ | 1174/1230 [8:52:42<25:38, 27.47s/it]
 96%|███████████████████████████████████▎ | 1175/1230 [8:53:07<24:36, 26.84s/it]
                                                                                
{'loss': 0.6959, 'grad_norm': 0.4331216514110565, 'learning_rate': 1.8877680202960894e-05, 'memory/max_active (GiB)': 8.37, 'memory/max_allocated (GiB)': 8.37, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 177.12, 'epoch': 0.83}

 96%|███████████████████████████████████▎ | 1175/1230 [8:53:07<24:36, 26.84s/it]
 96%|███████████████████████████████████▍ | 1176/1230 [8:53:32<23:41, 26.33s/it]
                                                                                
{'loss': 0.6651, 'grad_norm': 0.39659661054611206, 'learning_rate': 1.8846668589831026e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 179.17, 'epoch': 0.83}

 96%|███████████████████████████████████▍ | 1176/1230 [8:53:32<23:41, 26.33s/it]
 96%|███████████████████████████████████▍ | 1177/1230 [8:53:57<22:56, 25.96s/it]
                                                                                
{'loss': 0.6748, 'grad_norm': 0.4067908227443695, 'learning_rate': 1.8816211873754876e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.17, 'epoch': 0.83}

 96%|███████████████████████████████████▍ | 1177/1230 [8:53:57<22:56, 25.96s/it]
 96%|███████████████████████████████████▍ | 1178/1230 [8:54:23<22:21, 25.80s/it]
                                                                                
{'loss': 0.6775, 'grad_norm': 0.42076539993286133, 'learning_rate': 1.8786310265582577e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.67, 'epoch': 0.83}

 96%|███████████████████████████████████▍ | 1178/1230 [8:54:23<22:21, 25.80s/it]
 96%|███████████████████████████████████▍ | 1179/1230 [8:54:52<22:52, 26.91s/it]
                                                                                
{'loss': 0.6833, 'grad_norm': 0.38905006647109985, 'learning_rate': 1.8756963972321327e-05, 'memory/max_active (GiB)': 11.14, 'memory/max_allocated (GiB)': 11.14, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.36, 'epoch': 0.83}

 96%|███████████████████████████████████▍ | 1179/1230 [8:54:53<22:52, 26.91s/it]
 96%|███████████████████████████████████▍ | 1180/1230 [8:55:18<22:04, 26.49s/it]
                                                                                
{'loss': 0.6814, 'grad_norm': 0.4222973585128784, 'learning_rate': 1.872817319713384e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.35, 'epoch': 0.84}

 96%|███████████████████████████████████▍ | 1180/1230 [8:55:18<22:04, 26.49s/it]
 96%|███████████████████████████████████▌ | 1181/1230 [8:55:47<22:18, 27.32s/it]
                                                                                
{'loss': 0.7123, 'grad_norm': 0.42359399795532227, 'learning_rate': 1.8699938139337087e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 155.69, 'epoch': 0.84}

 96%|███████████████████████████████████▌ | 1181/1230 [8:55:47<22:18, 27.32s/it]
 96%|███████████████████████████████████▌ | 1182/1230 [8:56:13<21:33, 26.95s/it]
                                                                                
{'loss': 0.7042, 'grad_norm': 0.4278631806373596, 'learning_rate': 1.867225899440078e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.92, 'epoch': 0.84}

 96%|███████████████████████████████████▌ | 1182/1230 [8:56:13<21:33, 26.95s/it]
 96%|███████████████████████████████████▌ | 1183/1230 [8:56:42<21:39, 27.66s/it]
                                                                                
{'loss': 0.738, 'grad_norm': 0.4205786883831024, 'learning_rate': 1.8645135953946115e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.22, 'epoch': 0.84}

 96%|███████████████████████████████████▌ | 1183/1230 [8:56:42<21:39, 27.66s/it]
 96%|███████████████████████████████████▌ | 1184/1230 [8:57:08<20:40, 26.97s/it]
                                                                                
{'loss': 0.6726, 'grad_norm': 0.41535499691963196, 'learning_rate': 1.861856920574439e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.52, 'epoch': 0.84}

 96%|███████████████████████████████████▌ | 1184/1230 [8:57:08<20:40, 26.97s/it]
 96%|███████████████████████████████████▋ | 1185/1230 [8:57:34<20:01, 26.69s/it]
                                                                                
{'loss': 0.6825, 'grad_norm': 0.4117220342159271, 'learning_rate': 1.859255893371571e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.72, 'epoch': 0.84}

 96%|███████████████████████████████████▋ | 1185/1230 [8:57:34<20:01, 26.69s/it]
 96%|███████████████████████████████████▋ | 1186/1230 [8:58:03<20:11, 27.54s/it]
                                                                                
{'loss': 0.6929, 'grad_norm': 0.41078293323516846, 'learning_rate': 1.8567105317927763e-05, 'memory/max_active (GiB)': 12.16, 'memory/max_allocated (GiB)': 12.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.36, 'epoch': 0.84}

 96%|███████████████████████████████████▋ | 1186/1230 [8:58:03<20:11, 27.54s/it]
 97%|███████████████████████████████████▋ | 1187/1230 [8:58:29<19:18, 26.94s/it]
                                                                                
{'loss': 0.6941, 'grad_norm': 0.3880214989185333, 'learning_rate': 1.8542208534594493e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.81, 'epoch': 0.84}

 97%|███████████████████████████████████▋ | 1187/1230 [8:58:29<19:18, 26.94s/it]
 97%|███████████████████████████████████▋ | 1188/1230 [8:58:54<18:30, 26.44s/it]
                                                                                
{'loss': 0.6717, 'grad_norm': 0.42159828543663025, 'learning_rate': 1.8517868756074965e-05, 'memory/max_active (GiB)': 7.8, 'memory/max_allocated (GiB)': 7.8, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.74, 'epoch': 0.84}

 97%|███████████████████████████████████▋ | 1188/1230 [8:58:54<18:30, 26.44s/it]
 97%|███████████████████████████████████▊ | 1189/1230 [8:59:21<18:03, 26.44s/it]
                                                                                
{'loss': 0.6866, 'grad_norm': 0.4262861907482147, 'learning_rate': 1.8494086150872094e-05, 'memory/max_active (GiB)': 13.18, 'memory/max_allocated (GiB)': 13.18, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.37, 'epoch': 0.84}

 97%|███████████████████████████████████▊ | 1189/1230 [8:59:21<18:03, 26.44s/it]
 97%|███████████████████████████████████▊ | 1190/1230 [8:59:50<18:10, 27.26s/it]
                                                                                
{'loss': 0.6831, 'grad_norm': 0.4111368954181671, 'learning_rate': 1.847086088363154e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.54, 'epoch': 0.84}

 97%|███████████████████████████████████▊ | 1190/1230 [8:59:50<18:10, 27.26s/it]
 97%|███████████████████████████████████▊ | 1191/1230 [9:00:19<18:07, 27.87s/it]
                                                                                
{'loss': 0.6912, 'grad_norm': 0.4007345736026764, 'learning_rate': 1.8448193115140544e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.0, 'epoch': 0.84}

 97%|███████████████████████████████████▊ | 1191/1230 [9:00:19<18:07, 27.87s/it]
 97%|███████████████████████████████████▊ | 1192/1230 [9:00:48<17:56, 28.32s/it]
                                                                                
{'loss': 0.6632, 'grad_norm': 0.40354427695274353, 'learning_rate': 1.8426083002326793e-05, 'memory/max_active (GiB)': 12.92, 'memory/max_allocated (GiB)': 12.92, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 165.25, 'epoch': 0.84}

 97%|███████████████████████████████████▊ | 1192/1230 [9:00:48<17:56, 28.32s/it]
 97%|███████████████████████████████████▉ | 1193/1230 [9:01:15<17:03, 27.67s/it]
                                                                                
{'loss': 0.6789, 'grad_norm': 0.41621047258377075, 'learning_rate': 1.840453069825737e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 180.88, 'epoch': 0.84}

 97%|███████████████████████████████████▉ | 1193/1230 [9:01:15<17:03, 27.67s/it]
 97%|███████████████████████████████████▉ | 1194/1230 [9:01:44<16:55, 28.21s/it]
                                                                                
{'loss': 0.6742, 'grad_norm': 0.39938268065452576, 'learning_rate': 1.8383536352137688e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 162.04, 'epoch': 0.85}

 97%|███████████████████████████████████▉ | 1194/1230 [9:01:44<16:55, 28.21s/it]
 97%|███████████████████████████████████▉ | 1195/1230 [9:02:14<16:41, 28.61s/it]
                                                                                
{'loss': 0.6859, 'grad_norm': 0.38830655813217163, 'learning_rate': 1.836310010931043e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 170.31, 'epoch': 0.85}

 97%|███████████████████████████████████▉ | 1195/1230 [9:02:14<16:41, 28.61s/it]
 97%|███████████████████████████████████▉ | 1196/1230 [9:02:43<16:25, 28.98s/it]
                                                                                
{'loss': 0.6578, 'grad_norm': 0.40406686067581177, 'learning_rate': 1.8343222111254557e-05, 'memory/max_active (GiB)': 7.61, 'memory/max_allocated (GiB)': 7.61, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.22, 'epoch': 0.85}

 97%|███████████████████████████████████▉ | 1196/1230 [9:02:44<16:25, 28.98s/it]
 97%|████████████████████████████████████ | 1197/1230 [9:03:13<16:03, 29.19s/it]
                                                                                
{'loss': 0.6397, 'grad_norm': 0.404873251914978, 'learning_rate': 1.8323902495584362e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 158.96, 'epoch': 0.85}

 97%|████████████████████████████████████ | 1197/1230 [9:03:13<16:03, 29.19s/it]
 97%|████████████████████████████████████ | 1198/1230 [9:03:39<14:57, 28.05s/it]
                                                                                
{'loss': 0.696, 'grad_norm': 0.3995179533958435, 'learning_rate': 1.8305141396048456e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.35, 'epoch': 0.85}

 97%|████████████████████████████████████ | 1198/1230 [9:03:39<14:57, 28.05s/it]
 97%|████████████████████████████████████ | 1199/1230 [9:04:04<14:08, 27.37s/it]
                                                                                
{'loss': 0.7303, 'grad_norm': 0.42433929443359375, 'learning_rate': 1.82869389425289e-05, 'memory/max_active (GiB)': 11.66, 'memory/max_allocated (GiB)': 11.66, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.54, 'epoch': 0.85}

 97%|████████████████████████████████████ | 1199/1230 [9:04:04<14:08, 27.37s/it]
 98%|████████████████████████████████████ | 1200/1230 [9:04:29<13:20, 26.68s/it]
                                                                                
{'loss': 0.7329, 'grad_norm': 0.44642794132232666, 'learning_rate': 1.8269295261040258e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.33, 'epoch': 0.85}

 98%|████████████████████████████████████ | 1200/1230 [9:04:29<13:20, 26.68s/it]
 98%|████████████████████████████████████▏| 1201/1230 [9:04:54<12:37, 26.13s/it]
                                                                                
{'loss': 0.6645, 'grad_norm': 0.3976913392543793, 'learning_rate': 1.825221047372875e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.63, 'epoch': 0.85}

 98%|████████████████████████████████████▏| 1201/1230 [9:04:54<12:37, 26.13s/it]
 98%|████████████████████████████████████▏| 1202/1230 [9:05:24<12:37, 27.07s/it]
                                                                                
{'loss': 0.687, 'grad_norm': 0.42537352442741394, 'learning_rate': 1.8235684698871423e-05, 'memory/max_active (GiB)': 10.52, 'memory/max_allocated (GiB)': 10.52, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 143.15, 'epoch': 0.85}

 98%|████████████████████████████████████▏| 1202/1230 [9:05:24<12:37, 27.07s/it]
 98%|████████████████████████████████████▏| 1203/1230 [9:05:53<12:29, 27.74s/it]
                                                                                
{'loss': 0.6535, 'grad_norm': 0.3854252099990845, 'learning_rate': 1.8219718050875293e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 173.34, 'epoch': 0.85}

 98%|████████████████████████████████████▏| 1203/1230 [9:05:53<12:29, 27.74s/it]
 98%|████████████████████████████████████▏| 1204/1230 [9:06:18<11:41, 26.99s/it]
                                                                                
{'loss': 0.7002, 'grad_norm': 0.4698829650878906, 'learning_rate': 1.8204310640276564e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 172.25, 'epoch': 0.85}

 98%|████████████████████████████████████▏| 1204/1230 [9:06:18<11:41, 26.99s/it]
 98%|████████████████████████████████████▏| 1205/1230 [9:06:43<11:01, 26.48s/it]
                                                                                
{'loss': 0.6692, 'grad_norm': 0.4057368338108063, 'learning_rate': 1.818946257373989e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 190.63, 'epoch': 0.85}

 98%|████████████████████████████████████▏| 1205/1230 [9:06:43<11:01, 26.48s/it]
 98%|████████████████████████████████████▎| 1206/1230 [9:07:13<10:58, 27.44s/it]
                                                                                
{'loss': 0.6463, 'grad_norm': 0.40984222292900085, 'learning_rate': 1.817517395405761e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 154.15, 'epoch': 0.85}

 98%|████████████████████████████████████▎| 1206/1230 [9:07:13<10:58, 27.44s/it]
 98%|████████████████████████████████████▎| 1207/1230 [9:07:43<10:45, 28.08s/it]
                                                                                
{'loss': 0.7006, 'grad_norm': 0.4265860319137573, 'learning_rate': 1.8161444880149034e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 153.68, 'epoch': 0.85}

 98%|████████████████████████████████████▎| 1207/1230 [9:07:43<10:45, 28.08s/it]
 98%|████████████████████████████████████▎| 1208/1230 [9:08:08<10:02, 27.39s/it]
                                                                                
{'loss': 0.6773, 'grad_norm': 0.4249769151210785, 'learning_rate': 1.8148275447059797e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.19, 'epoch': 0.85}

 98%|████████████████████████████████████▎| 1208/1230 [9:08:08<10:02, 27.39s/it]
 98%|████████████████████████████████████▎| 1209/1230 [9:08:34<09:24, 26.90s/it]
                                                                                
{'loss': 0.6814, 'grad_norm': 0.4423636794090271, 'learning_rate': 1.813566574596114e-05, 'memory/max_active (GiB)': 10.64, 'memory/max_allocated (GiB)': 10.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 194.04, 'epoch': 0.86}

 98%|████████████████████████████████████▎| 1209/1230 [9:08:34<09:24, 26.90s/it]
 98%|████████████████████████████████████▍| 1210/1230 [9:09:04<09:13, 27.69s/it]
                                                                                
{'loss': 0.6896, 'grad_norm': 0.4211917817592621, 'learning_rate': 1.812361586414932e-05, 'memory/max_active (GiB)': 9.0, 'memory/max_allocated (GiB)': 9.0, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 149.04, 'epoch': 0.86}

 98%|████████████████████████████████████▍| 1210/1230 [9:09:04<09:13, 27.69s/it]
 98%|████████████████████████████████████▍| 1211/1230 [9:09:30<08:37, 27.22s/it]
                                                                                
{'loss': 0.6342, 'grad_norm': 0.40407997369766235, 'learning_rate': 1.811212588504499e-05, 'memory/max_active (GiB)': 12.67, 'memory/max_allocated (GiB)': 12.67, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 186.93, 'epoch': 0.86}

 98%|████████████████████████████████████▍| 1211/1230 [9:09:30<08:37, 27.22s/it]
 99%|████████████████████████████████████▍| 1212/1230 [9:09:55<07:58, 26.60s/it]
                                                                                
{'loss': 0.6884, 'grad_norm': 0.4168745279312134, 'learning_rate': 1.8101195888192665e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 195.37, 'epoch': 0.86}

 99%|████████████████████████████████████▍| 1212/1230 [9:09:55<07:58, 26.60s/it]
 99%|████████████████████████████████████▍| 1213/1230 [9:10:24<07:46, 27.47s/it]
                                                                                
{'loss': 0.6657, 'grad_norm': 0.4092898964881897, 'learning_rate': 1.8090825949260074e-05, 'memory/max_active (GiB)': 8.62, 'memory/max_allocated (GiB)': 8.62, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 152.63, 'epoch': 0.86}

 99%|████████████████████████████████████▍| 1213/1230 [9:10:24<07:46, 27.47s/it]
 99%|████████████████████████████████████▌| 1214/1230 [9:10:50<07:08, 26.80s/it]
                                                                                
{'loss': 0.693, 'grad_norm': 0.4300343990325928, 'learning_rate': 1.808101614003774e-05, 'memory/max_active (GiB)': 10.27, 'memory/max_allocated (GiB)': 10.27, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 171.55, 'epoch': 0.86}

 99%|████████████████████████████████████▌| 1214/1230 [9:10:50<07:08, 26.80s/it]
 99%|████████████████████████████████████▌| 1215/1230 [9:11:16<06:38, 26.56s/it]
                                                                                
{'loss': 0.6622, 'grad_norm': 0.40447312593460083, 'learning_rate': 1.8071766528438423e-05, 'memory/max_active (GiB)': 11.91, 'memory/max_allocated (GiB)': 11.91, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 196.53, 'epoch': 0.86}

 99%|████████████████████████████████████▌| 1215/1230 [9:11:16<06:38, 26.56s/it]
 99%|████████████████████████████████████▌| 1216/1230 [9:11:45<06:24, 27.46s/it]
                                                                                
{'loss': 0.7004, 'grad_norm': 0.39810648560523987, 'learning_rate': 1.8063077178496673e-05, 'memory/max_active (GiB)': 8.42, 'memory/max_allocated (GiB)': 8.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 181.81, 'epoch': 0.86}

 99%|████████████████████████████████████▌| 1216/1230 [9:11:45<06:24, 27.46s/it]
 99%|████████████████████████████████████▌| 1217/1230 [9:12:12<05:52, 27.10s/it]
                                                                                
{'loss': 0.6561, 'grad_norm': 0.399292916059494, 'learning_rate': 1.8054948150368386e-05, 'memory/max_active (GiB)': 12.93, 'memory/max_allocated (GiB)': 12.93, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 184.91, 'epoch': 0.86}

 99%|████████████████████████████████████▌| 1217/1230 [9:12:12<05:52, 27.10s/it]
 99%|████████████████████████████████████▋| 1218/1230 [9:12:41<05:33, 27.77s/it]
                                                                                
{'loss': 0.6716, 'grad_norm': 0.42511075735092163, 'learning_rate': 1.804737950033035e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 156.87, 'epoch': 0.86}

 99%|████████████████████████████████████▋| 1218/1230 [9:12:41<05:33, 27.77s/it]
 99%|████████████████████████████████████▋| 1219/1230 [9:13:10<05:10, 28.26s/it]
                                                                                
{'loss': 0.697, 'grad_norm': 0.41346096992492676, 'learning_rate': 1.804037128077992e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.92, 'epoch': 0.86}

 99%|████████████████████████████████████▋| 1219/1230 [9:13:10<05:10, 28.26s/it]
 99%|████████████████████████████████████▋| 1220/1230 [9:13:35<04:33, 27.33s/it]
                                                                                
{'loss': 0.674, 'grad_norm': 0.4128546416759491, 'learning_rate': 1.803392354023462e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 183.29, 'epoch': 0.86}

 99%|████████████████████████████████████▋| 1220/1230 [9:13:35<04:33, 27.33s/it]
 99%|████████████████████████████████████▋| 1221/1230 [9:14:05<04:12, 28.02s/it]
                                                                                
{'loss': 0.7091, 'grad_norm': 0.41580066084861755, 'learning_rate': 1.8028036323331777e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 157.53, 'epoch': 0.86}

 99%|████████████████████████████████████▋| 1221/1230 [9:14:05<04:12, 28.02s/it]
 99%|████████████████████████████████████▊| 1222/1230 [9:14:34<03:46, 28.36s/it]
                                                                                
{'loss': 0.6831, 'grad_norm': 0.38868921995162964, 'learning_rate': 1.802270967082829e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 167.84, 'epoch': 0.86}

 99%|████████████████████████████████████▊| 1222/1230 [9:14:34<03:46, 28.36s/it]
 99%|████████████████████████████████████▊| 1223/1230 [9:15:00<03:12, 27.46s/it]
                                                                                
{'loss': 0.6983, 'grad_norm': 0.42300495505332947, 'learning_rate': 1.8017943619600247e-05, 'memory/max_active (GiB)': 7.35, 'memory/max_allocated (GiB)': 7.35, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 182.1, 'epoch': 0.87}

 99%|████████████████████████████████████▊| 1223/1230 [9:15:00<03:12, 27.46s/it]
100%|████████████████████████████████████▊| 1224/1230 [9:15:25<02:40, 26.80s/it]
                                                                                
{'loss': 0.6763, 'grad_norm': 0.39557501673698425, 'learning_rate': 1.8013738202642787e-05, 'memory/max_active (GiB)': 7.99, 'memory/max_allocated (GiB)': 7.99, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 178.0, 'epoch': 0.87}

100%|████████████████████████████████████▊| 1224/1230 [9:15:25<02:40, 26.80s/it]
100%|████████████████████████████████████▊| 1225/1230 [9:15:54<02:18, 27.62s/it]
                                                                                
{'loss': 0.6488, 'grad_norm': 0.37639757990837097, 'learning_rate': 1.801009344906976e-05, 'memory/max_active (GiB)': 12.42, 'memory/max_allocated (GiB)': 12.42, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 168.59, 'epoch': 0.87}

100%|████████████████████████████████████▊| 1225/1230 [9:15:54<02:18, 27.62s/it]
100%|████████████████████████████████████▉| 1226/1230 [9:16:24<01:52, 28.21s/it]
                                                                                
{'loss': 0.6913, 'grad_norm': 0.40694984793663025, 'learning_rate': 1.8007009384113606e-05, 'memory/max_active (GiB)': 11.4, 'memory/max_allocated (GiB)': 11.4, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 166.92, 'epoch': 0.87}

100%|████████████████████████████████████▉| 1226/1230 [9:16:24<01:52, 28.21s/it]
100%|████████████████████████████████████▉| 1227/1230 [9:16:49<01:22, 27.37s/it]
                                                                                
{'loss': 0.6735, 'grad_norm': 0.4052731394767761, 'learning_rate': 1.8004486029125108e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 193.4, 'epoch': 0.87}

100%|████████████████████████████████████▉| 1227/1230 [9:16:49<01:22, 27.37s/it]
100%|████████████████████████████████████▉| 1228/1230 [9:17:19<00:55, 27.94s/it]
                                                                                
{'loss': 0.6655, 'grad_norm': 0.41497644782066345, 'learning_rate': 1.800252340157334e-05, 'memory/max_active (GiB)': 7.17, 'memory/max_allocated (GiB)': 7.17, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 159.07, 'epoch': 0.87}

100%|████████████████████████████████████▉| 1228/1230 [9:17:19<00:55, 27.94s/it]
100%|████████████████████████████████████▉| 1229/1230 [9:17:48<00:28, 28.31s/it]
                                                                                
{'loss': 0.7009, 'grad_norm': 0.4352482557296753, 'learning_rate': 1.8001121515045448e-05, 'memory/max_active (GiB)': 7.16, 'memory/max_allocated (GiB)': 7.16, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 163.05, 'epoch': 0.87}

100%|████████████████████████████████████▉| 1229/1230 [9:17:48<00:28, 28.31s/it]
100%|█████████████████████████████████████| 1230/1230 [9:18:13<00:00, 27.41s/it]
                                                                                
{'loss': 0.6811, 'grad_norm': 0.421597421169281, 'learning_rate': 1.800028037924663e-05, 'memory/max_active (GiB)': 9.64, 'memory/max_allocated (GiB)': 9.64, 'memory/device_reserved (GiB)': 13.43, 'tokens_per_second_per_gpu': 176.99, 'epoch': 0.87}

100%|█████████████████████████████████████| 1230/1230 [9:18:13<00:00, 27.41s/it][2025-10-21 12:26:45,011] [INFO] [axolotl.core.trainers.base._save:664] [PID:687] Saving model checkpoint to /kaggle/working/outputs/qwen-sft-mt-kmvi/checkpoint-1230

                                                                                
{'train_runtime': 33497.2495, 'train_samples_per_second': 0.588, 'train_steps_per_second': 0.037, 'train_loss': 0.8053028226383333, 'memory/max_active (GiB)': 1.62, 'memory/max_allocated (GiB)': 1.62, 'memory/device_reserved (GiB)': 13.43, 'epoch': 0.87}

100%|█████████████████████████████████████| 1230/1230 [9:18:14<00:00, 27.41s/it]
100%|█████████████████████████████████████| 1230/1230 [9:18:14<00:00, 27.23s/it]
[2025-10-21 12:26:46,793] [INFO] [axolotl.train.save_trained_model:218] [PID:687] Training completed! Saving trained model to /kaggle/working/outputs/qwen-sft-mt-kmvi.
[2025-10-21 12:26:47,227] [INFO] [axolotl.train.save_trained_model:336] [PID:687] Model successfully saved to /kaggle/working/outputs/qwen-sft-mt-kmvi