[2025-07-06 10:18:37,401][__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: microsoft/phi-4 type: phi4_classification_lora num_labels: 6 output_dir: ./results/phi35-balanced/C4 logging_dir: ./logs/phi35-balanced/C4 best_model_dir: ./results/phi35-balanced/C4/best_model lora_r: 16 lora_dropout: 0.1 lora_alpha: 32 lora_target_modules: all-linear tokenizer: name: microsoft/phi-4 dataset: grade_index: 3 use_full_context: true training_params: weight_decay: 0.01 warmup_ratio: 0.1 learning_rate: 5.0e-05 train_batch_size: 8 eval_batch_size: 4 gradient_accumulation_steps: 2 gradient_checkpointing: true [2025-07-06 10:18:41,292][__main__][INFO] - GPU 0: NVIDIA H200 | TDP ≈ 700 W [2025-07-06 10:18:41,292][__main__][INFO] - Starting the Fine Tuning training process. [2025-07-06 10:18:45,062][transformers.tokenization_utils_base][INFO] - loading file vocab.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/vocab.json [2025-07-06 10:18:45,062][transformers.tokenization_utils_base][INFO] - loading file merges.txt from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/merges.txt [2025-07-06 10:18:45,063][transformers.tokenization_utils_base][INFO] - loading file tokenizer.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/tokenizer.json [2025-07-06 10:18:45,063][transformers.tokenization_utils_base][INFO] - loading file added_tokens.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/added_tokens.json [2025-07-06 10:18:45,063][transformers.tokenization_utils_base][INFO] - loading file special_tokens_map.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/special_tokens_map.json [2025-07-06 10:18:45,063][transformers.tokenization_utils_base][INFO] - loading file tokenizer_config.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/tokenizer_config.json [2025-07-06 10:18:45,063][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None [2025-07-06 10:18:45,212][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False; Use Full Context: True [2025-07-06 10:18:45,684][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 10:18:45,685][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "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": 17920, "label2id": { "LABEL_0": 0, "LABEL_1": 1, "LABEL_2": 2, "LABEL_3": 3, "LABEL_4": 4, "LABEL_5": 5 }, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 10:18:45,814][transformers.modeling_utils][INFO] - loading weights file model.safetensors from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/model.safetensors.index.json [2025-07-06 10:18:45,814][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object [2025-07-06 10:18:45,814][transformers.modeling_utils][INFO] - Instantiating Phi3ForSequenceClassification model under default dtype torch.bfloat16. [2025-07-06 10:19:03,781][transformers.modeling_utils][INFO] - Some weights of the model checkpoint at microsoft/phi-4 were not used when initializing Phi3ForSequenceClassification: ['lm_head.weight'] - This IS expected if you are initializing Phi3ForSequenceClassification 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 Phi3ForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model). [2025-07-06 10:19:03,781][transformers.modeling_utils][WARNING] - Some weights of Phi3ForSequenceClassification were not initialized from the model checkpoint at microsoft/phi-4 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. [2025-07-06 10:19:06,343][__main__][INFO] - Initialized new PEFT model for ce loss [2025-07-06 10:19:06,346][__main__][INFO] - None [2025-07-06 10:19:06,347][transformers.training_args][INFO] - PyTorch: setting up devices [2025-07-06 10:19:06,422][__main__][INFO] - Total steps: 620. Number of warmup steps: 62 [2025-07-06 10:19:06,429][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. [2025-07-06 10:19:06,464][transformers.trainer][INFO] - Using auto half precision backend [2025-07-06 10:19:06,465][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. [2025-07-06 10:19:06,467][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:19:06,480][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 10:19:06,480][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 10:19:06,480][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 10:19:46,263][transformers.trainer][INFO] - The following columns in the Training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:19:46,303][transformers.trainer][INFO] - ***** Running training ***** [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Num examples = 500 [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Num Epochs = 20 [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Instantaneous batch size per device = 8 [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Total train batch size (w. parallel, distributed & accumulation) = 16 [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Gradient Accumulation steps = 2 [2025-07-06 10:19:46,303][transformers.trainer][INFO] - Total optimization steps = 640 [2025-07-06 10:19:46,305][transformers.trainer][INFO] - Number of trainable parameters = 55,736,320 [2025-07-06 10:19:46,386][transformers.models.phi3.modeling_phi3][WARNING] - `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. [2025-07-06 10:28:03,700][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:28:03,703][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 10:28:03,703][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 10:28:03,703][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 10:28:43,290][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-32 [2025-07-06 10:28:43,823][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 10:28:43,824][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 10:37:02,466][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:37:02,470][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 10:37:02,470][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 10:37:02,470][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 10:37:41,978][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-64 [2025-07-06 10:37:42,310][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 10:37:42,311][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 10:46:02,560][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:46:02,564][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 10:46:02,564][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 10:46:02,564][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 10:46:42,107][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-96 [2025-07-06 10:46:42,454][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 10:46:42,455][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 10:46:43,497][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-64] due to args.save_total_limit [2025-07-06 10:55:01,135][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 10:55:01,138][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 10:55:01,138][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 10:55:01,138][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 10:55:40,666][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-128 [2025-07-06 10:55:41,013][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 10:55:41,014][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 10:55:42,085][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-32] due to args.save_total_limit [2025-07-06 10:55:42,121][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-96] due to args.save_total_limit [2025-07-06 11:03:59,827][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:03:59,831][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:03:59,831][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:03:59,831][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:04:39,335][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-160 [2025-07-06 11:04:39,660][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:04:39,661][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:04:40,782][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-128] due to args.save_total_limit [2025-07-06 11:12:58,358][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:12:58,362][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:12:58,362][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:12:58,362][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:13:37,889][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-192 [2025-07-06 11:13:38,236][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:13:38,237][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:21:57,070][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:21:57,074][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:21:57,074][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:21:57,074][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:22:36,599][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-224 [2025-07-06 11:22:36,932][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:22:36,933][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:22:38,082][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-192] due to args.save_total_limit [2025-07-06 11:30:55,565][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:30:55,569][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:30:55,569][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:30:55,569][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:31:35,125][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-256 [2025-07-06 11:31:35,609][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:31:35,610][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:31:36,694][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-160] due to args.save_total_limit [2025-07-06 11:31:36,730][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-224] due to args.save_total_limit [2025-07-06 11:39:54,595][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:39:54,599][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:39:54,599][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:39:54,599][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:40:34,144][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-288 [2025-07-06 11:40:34,482][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:40:34,483][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:48:53,446][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:48:53,450][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:48:53,450][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:48:53,450][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:49:32,972][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-320 [2025-07-06 11:49:33,472][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:49:33,473][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:49:34,678][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-288] due to args.save_total_limit [2025-07-06 11:57:52,975][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 11:57:52,979][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 11:57:52,979][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 11:57:52,979][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 11:58:32,483][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-352 [2025-07-06 11:58:32,934][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 11:58:32,935][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 11:58:34,219][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-320] due to args.save_total_limit [2025-07-06 12:06:52,237][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 12:06:52,241][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 12:06:52,241][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 12:06:52,241][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 12:07:31,904][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-384 [2025-07-06 12:07:32,244][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 12:07:32,245][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 12:07:33,387][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-352] due to args.save_total_limit [2025-07-06 12:15:51,391][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 12:15:51,395][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 12:15:51,395][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 12:15:51,395][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 12:16:30,969][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-416 [2025-07-06 12:16:31,291][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 12:16:31,292][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 12:16:32,367][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-384] due to args.save_total_limit [2025-07-06 12:16:32,401][transformers.trainer][INFO] - Training completed. Do not forget to share your model on huggingface.co/models =) [2025-07-06 12:16:32,401][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-256 (score: 0.5432363013698631). [2025-07-06 12:16:32,668][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/10-18-37/results/phi35-balanced/C4/checkpoint-416] due to args.save_total_limit [2025-07-06 12:16:32,704][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 12:16:32,713][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 12:16:32,713][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 12:16:32,713][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 12:17:12,251][__main__][INFO] - Training completed successfully. [2025-07-06 12:17:12,251][__main__][INFO] - Running on Test [2025-07-06 12:17:12,252][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt. If supporting_text, grades, id, essay_text, id_prompt, reference, essay_year, prompt are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 12:17:12,254][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 12:17:12,255][transformers.trainer][INFO] - Num examples = 138 [2025-07-06 12:17:12,255][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 12:17:54,329][__main__][INFO] - Test metrics: {'eval_loss': 1.1753787994384766, 'eval_model_preparation_time': 0.0096, 'eval_accuracy': 0.4782608695652174, 'eval_RMSE': 31.20757990421976, 'eval_QWK': 0.5983367983367983, 'eval_HDIV': 0.0, 'eval_Macro_F1': 0.28940643739961425, 'eval_Micro_F1': 0.4782608695652174, 'eval_Weighted_F1': 0.49833793115813957, 'eval_TP_0': 0, 'eval_TN_0': 137, 'eval_FP_0': 0, 'eval_FN_0': 1, 'eval_TP_1': 0, 'eval_TN_1': 137, 'eval_FP_1': 0, 'eval_FN_1': 1, 'eval_TP_2': 7, 'eval_TN_2': 104, 'eval_FP_2': 25, 'eval_FN_2': 2, 'eval_TP_3': 30, 'eval_TN_3': 54, 'eval_FP_3': 8, 'eval_FN_3': 46, 'eval_TP_4': 25, 'eval_TN_4': 67, 'eval_FP_4': 25, 'eval_FN_4': 21, 'eval_TP_5': 4, 'eval_TN_5': 119, 'eval_FP_5': 14, 'eval_FN_5': 1, 'eval_runtime': 42.0636, 'eval_samples_per_second': 3.281, 'eval_steps_per_second': 0.832, 'epoch': 13.0} [2025-07-06 12:17:54,329][transformers.trainer][INFO] - Saving model checkpoint to ./results/phi35-balanced/C4/best_model [2025-07-06 12:17:54,656][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /workspace/.hf_home/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 12:17:54,657][transformers.configuration_utils][INFO] - Model config Phi3Config { "architectures": [ "Phi3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 100257, "embd_pdrop": 0.0, "eos_token_id": 100265, "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 16384, "model_type": "phi3", "num_attention_heads": 40, "num_hidden_layers": 40, "num_key_value_heads": 10, "original_max_position_embeddings": 16384, "pad_token_id": 100349, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 250000, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.53.0", "use_cache": true, "vocab_size": 100352 } [2025-07-06 12:17:55,417][transformers.tokenization_utils_base][INFO] - chat template saved in ./results/phi35-balanced/C4/best_model/chat_template.jinja [2025-07-06 12:17:55,418][transformers.tokenization_utils_base][INFO] - tokenizer config file saved in ./results/phi35-balanced/C4/best_model/tokenizer_config.json [2025-07-06 12:17:55,418][transformers.tokenization_utils_base][INFO] - Special tokens file saved in ./results/phi35-balanced/C4/best_model/special_tokens_map.json [2025-07-06 12:17:55,491][__main__][INFO] - Model and tokenizer saved to ./results/phi35-balanced/C4/best_model [2025-07-06 12:17:55,508][__main__][INFO] - Fine Tuning Finished. [2025-07-06 12:17:56,017][__main__][INFO] - Total emissions: 0.0926 kg CO2eq