[2025-06-30 04:45:31,825][__main__][INFO] - cache_dir: /tmp/ dataset: name: kamel-usp/aes_enem_dataset split: JBCS2025 training_params: seed: 42 num_train_epochs: 20 logging_steps: 100 metric_for_best_model: QWK bf16: true bootstrap: enabled: true n_bootstrap: 10000 bootstrap_seed: 42 metrics: - QWK - Macro_F1 - Weighted_F1 post_training_results: model_path: /workspace/jbcs2025/outputs/2025-03-24/20-42-59 experiments: model: name: meta-llama/Llama-3.1-8B type: llama31_classification_lora num_labels: 6 output_dir: ./results/ logging_dir: ./logs/ best_model_dir: ./results/best_model lora_r: 8 lora_dropout: 0.05 lora_alpha: 16 lora_target_modules: all-linear checkpoint_path: '' tokenizer: name: meta-llama/Llama-3.1-8B dataset: grade_index: 4 use_full_context: true training_params: weight_decay: 0.01 warmup_ratio: 0.1 learning_rate: 5.0e-05 train_batch_size: 1 eval_batch_size: 16 gradient_accumulation_steps: 16 gradient_checkpointing: false [2025-06-30 04:45:35,862][__main__][INFO] - GPU 0: NVIDIA H200 | TDP ≈ 700 W [2025-06-30 04:45:35,862][__main__][INFO] - Starting the Fine Tuning training process. [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file tokenizer.json from cache at /tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/tokenizer.json [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file tokenizer.model from cache at None [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file added_tokens.json from cache at None [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file special_tokens_map.json from cache at /tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/special_tokens_map.json [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file tokenizer_config.json from cache at /tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/tokenizer_config.json [2025-06-30 04:45:41,370][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None [2025-06-30 04:45:41,640][transformers.tokenization_utils_base][INFO] - Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. [2025-06-30 04:45:41,701][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False; Use Full Context: True [2025-06-30 04:45:43,486][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 04:45:43,488][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "id2label": { "0": "LABEL_0", "1": "LABEL_1", "2": "LABEL_2", "3": "LABEL_3", "4": "LABEL_4", "5": "LABEL_5" }, "initializer_range": 0.02, "intermediate_size": 14336, "label2id": { "LABEL_0": 0, "LABEL_1": 1, "LABEL_2": 2, "LABEL_3": 3, "LABEL_4": 4, "LABEL_5": 5 }, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 04:45:43,631][transformers.modeling_utils][INFO] - loading weights file model.safetensors from cache at /tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/model.safetensors.index.json [2025-06-30 04:45:43,632][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object [2025-06-30 04:45:43,632][transformers.modeling_utils][INFO] - Instantiating LlamaForSequenceClassification model under default dtype torch.bfloat16. [2025-06-30 04:45:49,664][transformers.modeling_utils][INFO] - Some weights of the model checkpoint at meta-llama/Llama-3.1-8B were not used when initializing LlamaForSequenceClassification: ['lm_head.weight'] - This IS expected if you are initializing LlamaForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model). - This IS NOT expected if you are initializing LlamaForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model). [2025-06-30 04:45:49,664][transformers.modeling_utils][WARNING] - Some weights of LlamaForSequenceClassification were not initialized from the model checkpoint at meta-llama/Llama-3.1-8B and are newly initialized: ['score.weight'] You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. [2025-06-30 04:45:51,001][__main__][INFO] - Initialized new PEFT model for ce loss [2025-06-30 04:45:51,004][__main__][INFO] - None [2025-06-30 04:45:51,004][transformers.training_args][INFO] - PyTorch: setting up devices [2025-06-30 04:45:51,067][__main__][INFO] - Total steps: 620. Number of warmup steps: 62 [2025-06-30 04:45:51,073][transformers.trainer][INFO] - You have loaded a model on multiple GPUs. `is_model_parallel` attribute will be force-set to `True` to avoid any unexpected behavior such as device placement mismatching. [2025-06-30 04:45:51,097][transformers.trainer][INFO] - Using auto half precision backend [2025-06-30 04:45:51,097][transformers.trainer][WARNING] - No label_names provided for model class `PeftModelForSequenceClassification`. Since `PeftModel` hides base models input arguments, if label_names is not given, label_names can't be set automatically within `Trainer`. Note that empty label_names list will be used instead. [2025-06-30 04:45:51,099][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 04:45:51,115][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 04:45:51,115][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 04:45:51,115][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 04:46:13,862][transformers.trainer][INFO] - The following columns in the Training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 04:46:13,915][transformers.trainer][INFO] - ***** Running training ***** [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Num examples = 500 [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Num Epochs = 20 [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Instantaneous batch size per device = 1 [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Total train batch size (w. parallel, distributed & accumulation) = 16 [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Gradient Accumulation steps = 16 [2025-06-30 04:46:13,915][transformers.trainer][INFO] - Total optimization steps = 640 [2025-06-30 04:46:13,918][transformers.trainer][INFO] - Number of trainable parameters = 20,996,096 [2025-06-30 04:50:13,565][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 04:50:13,569][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 04:50:13,569][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 04:50:13,569][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 04:50:35,997][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-32 [2025-06-30 04:50:36,487][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 04:50:36,487][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 04:54:36,445][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 04:54:36,449][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 04:54:36,449][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 04:54:36,449][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 04:54:58,890][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-64 [2025-06-30 04:54:59,349][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 04:54:59,350][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 04:54:59,769][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-32] due to args.save_total_limit [2025-06-30 04:58:59,590][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 04:58:59,593][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 04:58:59,594][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 04:58:59,594][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 04:59:22,004][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-96 [2025-06-30 04:59:22,486][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 04:59:22,487][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 04:59:22,896][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-64] due to args.save_total_limit [2025-06-30 05:03:22,548][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:03:22,552][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:03:22,552][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:03:22,552][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:03:44,959][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-128 [2025-06-30 05:03:45,464][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:03:45,465][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:03:45,887][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-96] due to args.save_total_limit [2025-06-30 05:07:45,419][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:07:45,423][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:07:45,423][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:07:45,423][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:08:07,853][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-160 [2025-06-30 05:08:08,312][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:08:08,312][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:08:08,677][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-128] due to args.save_total_limit [2025-06-30 05:12:08,351][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:12:08,355][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:12:08,355][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:12:08,355][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:12:30,781][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-192 [2025-06-30 05:12:31,292][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:12:31,293][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:12:31,692][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-160] due to args.save_total_limit [2025-06-30 05:16:31,206][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:16:31,210][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:16:31,210][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:16:31,210][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:16:53,632][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-224 [2025-06-30 05:16:54,189][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:16:54,190][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:16:54,588][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-192] due to args.save_total_limit [2025-06-30 05:20:54,534][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:20:54,537][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:20:54,537][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:20:54,537][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:21:16,967][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-256 [2025-06-30 05:21:17,418][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:21:17,418][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:25:17,622][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:25:17,626][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:25:17,626][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:25:17,626][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:25:40,050][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-288 [2025-06-30 05:25:40,504][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:25:40,505][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:25:40,886][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-224] due to args.save_total_limit [2025-06-30 05:25:40,906][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-256] due to args.save_total_limit [2025-06-30 05:29:41,079][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:29:41,083][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:29:41,083][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:29:41,083][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:30:03,492][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-320 [2025-06-30 05:30:04,158][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:30:04,159][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:30:04,578][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-288] due to args.save_total_limit [2025-06-30 05:34:04,750][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:34:04,753][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:34:04,753][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:34:04,753][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:34:27,160][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-352 [2025-06-30 05:34:27,828][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:34:27,828][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:34:28,177][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-320] due to args.save_total_limit [2025-06-30 05:38:28,631][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:38:28,634][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:38:28,635][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:38:28,635][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:38:51,062][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-384 [2025-06-30 05:38:51,538][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:38:51,539][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:42:52,194][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:42:52,197][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:42:52,197][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:42:52,198][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:43:14,624][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-416 [2025-06-30 05:43:15,061][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:43:15,062][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:43:15,478][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-384] due to args.save_total_limit [2025-06-30 05:47:15,223][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:47:15,227][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:47:15,227][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:47:15,227][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:47:37,662][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-448 [2025-06-30 05:47:38,118][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:47:38,118][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:47:38,477][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-416] due to args.save_total_limit [2025-06-30 05:51:38,414][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:51:38,417][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:51:38,418][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:51:38,418][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:52:00,845][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-480 [2025-06-30 05:52:01,371][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:52:01,371][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:52:01,784][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-448] due to args.save_total_limit [2025-06-30 05:56:01,703][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:56:01,708][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:56:01,708][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:56:01,708][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:56:24,182][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-512 [2025-06-30 05:56:24,649][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:56:24,650][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:56:25,053][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-480] due to args.save_total_limit [2025-06-30 05:56:25,070][transformers.trainer][INFO] - Training completed. Do not forget to share your model on huggingface.co/models =) [2025-06-30 05:56:25,070][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-352 (score: 0.3633916554508748). [2025-06-30 05:56:25,279][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-06-30/04-45-31/results/checkpoint-512] due to args.save_total_limit [2025-06-30 05:56:25,299][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:56:25,303][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:56:25,303][transformers.trainer][INFO] - Num examples = 132 [2025-06-30 05:56:25,303][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:56:47,762][__main__][INFO] - Training completed successfully. [2025-06-30 05:56:47,762][__main__][INFO] - Running on Test [2025-06-30 05:56:47,762][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text. If id, essay_year, reference, supporting_text, prompt, grades, id_prompt, essay_text are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-06-30 05:56:47,766][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-06-30 05:56:47,766][transformers.trainer][INFO] - Num examples = 138 [2025-06-30 05:56:47,766][transformers.trainer][INFO] - Batch size = 16 [2025-06-30 05:57:11,531][__main__][INFO] - Test metrics: {'eval_loss': 2.1231963634490967, 'eval_model_preparation_time': 0.0124, 'eval_accuracy': 0.2391304347826087, 'eval_RMSE': 76.7453912131782, 'eval_QWK': 0.1835841058368659, 'eval_HDIV': 0.21739130434782605, 'eval_Macro_F1': 0.24872321614312462, 'eval_Micro_F1': 0.2391304347826087, 'eval_Weighted_F1': 0.22157700274056857, 'eval_TP_0': 7, 'eval_TN_0': 97, 'eval_FP_0': 19, 'eval_FN_0': 15, 'eval_TP_1': 2, 'eval_TN_1': 96, 'eval_FP_1': 10, 'eval_FN_1': 30, 'eval_TP_2': 2, 'eval_TN_2': 94, 'eval_FP_2': 20, 'eval_FN_2': 22, 'eval_TP_3': 8, 'eval_TN_3': 89, 'eval_FP_3': 24, 'eval_FN_3': 17, 'eval_TP_4': 13, 'eval_TN_4': 75, 'eval_FP_4': 31, 'eval_FN_4': 19, 'eval_TP_5': 1, 'eval_TN_5': 134, 'eval_FP_5': 1, 'eval_FN_5': 2, 'eval_runtime': 23.753, 'eval_samples_per_second': 5.81, 'eval_steps_per_second': 0.379, 'epoch': 16.0} [2025-06-30 05:57:11,532][transformers.trainer][INFO] - Saving model checkpoint to ./results/best_model [2025-06-30 05:57:11,959][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json [2025-06-30 05:57:11,960][transformers.configuration_utils][INFO] - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "eos_token_id": 128001, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 14336, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 8.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 128256 } [2025-06-30 05:57:12,206][transformers.tokenization_utils_base][INFO] - tokenizer config file saved in ./results/best_model/tokenizer_config.json [2025-06-30 05:57:12,206][transformers.tokenization_utils_base][INFO] - Special tokens file saved in ./results/best_model/special_tokens_map.json [2025-06-30 05:57:12,338][__main__][INFO] - Model and tokenizer saved to ./results/best_model [2025-06-30 05:57:12,371][__main__][INFO] - Fine Tuning Finished. [2025-06-30 05:57:12,880][__main__][INFO] - Total emissions: 0.4253 kg CO2eq