[2025-07-06 05:42:53,025][__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/phi4-balanced/C2 logging_dir: ./logs/phi4-balanced/C2 best_model_dir: ./results/phi4-balanced/C2/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: 1 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 05:42:56,955][__main__][INFO] - GPU 0: NVIDIA H200 | TDP ≈ 700 W [2025-07-06 05:42:56,955][__main__][INFO] - Starting the Fine Tuning training process. [2025-07-06 05:43:00,852][transformers.tokenization_utils_base][INFO] - loading file vocab.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/vocab.json [2025-07-06 05:43:00,852][transformers.tokenization_utils_base][INFO] - loading file merges.txt from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/merges.txt [2025-07-06 05:43:00,852][transformers.tokenization_utils_base][INFO] - loading file tokenizer.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/tokenizer.json [2025-07-06 05:43:00,852][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 05:43:00,852][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 05:43:00,852][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 05:43:00,852][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None [2025-07-06 05:43:00,992][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False; Use Full Context: True [2025-07-06 05:43:01,238][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json [2025-07-06 05:43:01,238][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 05:43:01,365][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 05:43:01,366][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object [2025-07-06 05:43:01,366][transformers.modeling_utils][INFO] - Instantiating Phi3ForSequenceClassification model under default dtype torch.bfloat16. [2025-07-06 05:43:19,829][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 05:43:19,829][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 05:43:21,857][__main__][INFO] - Initialized new PEFT model for ce loss [2025-07-06 05:43:21,860][__main__][INFO] - None [2025-07-06 05:43:21,861][transformers.training_args][INFO] - PyTorch: setting up devices [2025-07-06 05:43:21,900][__main__][INFO] - Total steps: 620. Number of warmup steps: 62 [2025-07-06 05:43:21,907][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 05:43:21,940][transformers.trainer][INFO] - Using auto half precision backend [2025-07-06 05:43:21,941][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 05:43:21,942][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 05:43:21,955][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 05:43:21,955][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 05:43:21,955][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 05:44:09,428][transformers.trainer][INFO] - The following columns in the Training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 05:44:09,469][transformers.trainer][INFO] - ***** Running training ***** [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Num examples = 500 [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Num Epochs = 20 [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Instantaneous batch size per device = 8 [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Total train batch size (w. parallel, distributed & accumulation) = 16 [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Gradient Accumulation steps = 2 [2025-07-06 05:44:09,469][transformers.trainer][INFO] - Total optimization steps = 640 [2025-07-06 05:44:09,471][transformers.trainer][INFO] - Number of trainable parameters = 55,736,320 [2025-07-06 05:44:09,568][transformers.models.phi3.modeling_phi3][WARNING] - `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. [2025-07-06 05:53:55,371][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 05:53:55,374][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 05:53:55,374][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 05:53:55,374][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 05:54:42,590][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-32 [2025-07-06 05:54:42,962][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 05:54:42,962][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 06:04:29,707][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:04:29,711][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:04:29,711][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:04:29,711][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:05:16,901][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-64 [2025-07-06 06:05:17,244][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 06:05:17,244][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 06:05:18,342][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-32] due to args.save_total_limit [2025-07-06 06:15:04,258][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:15:04,262][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:15:04,262][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:15:04,262][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:15:51,472][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-96 [2025-07-06 06:15:51,812][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 06:15:51,813][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 06:15:52,857][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-64] due to args.save_total_limit [2025-07-06 06:25:38,456][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:25:38,459][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:25:38,459][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:25:38,459][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:26:25,664][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-128 [2025-07-06 06:26:26,151][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 06:26:26,152][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 06:26:27,245][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-96] due to args.save_total_limit [2025-07-06 06:36:13,060][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:36:13,063][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:36:13,064][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:36:13,064][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:37:00,261][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-160 [2025-07-06 06:37:00,594][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 06:37:00,595][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 06:46:47,163][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:46:47,166][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:46:47,166][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:46:47,166][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:47:34,344][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-192 [2025-07-06 06:47:34,677][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 06:47:34,677][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 06:47:35,674][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-160] due to args.save_total_limit [2025-07-06 06:57:21,183][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 06:57:21,186][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 06:57:21,186][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 06:57:21,186][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 06:58:08,374][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-224 [2025-07-06 06:58:08,706][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 06:58:08,706][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 06:58:09,678][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-128] due to args.save_total_limit [2025-07-06 06:58:09,713][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-192] due to args.save_total_limit [2025-07-06 07:07:55,067][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 07:07:55,071][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 07:07:55,071][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 07:07:55,071][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 07:08:42,214][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-256 [2025-07-06 07:08:42,731][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 07:08:42,732][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 07:08:43,869][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-224] due to args.save_total_limit [2025-07-06 07:18:29,265][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 07:18:29,268][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 07:18:29,268][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 07:18:29,268][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 07:19:16,476][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-288 [2025-07-06 07:19:16,832][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 07:19:16,833][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 07:29:03,608][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 07:29:03,612][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 07:29:03,612][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 07:29:03,612][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 07:29:50,809][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-320 [2025-07-06 07:29:51,119][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 07:29:51,119][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 07:29:52,167][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-288] due to args.save_total_limit [2025-07-06 07:39:38,201][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 07:39:38,204][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 07:39:38,205][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 07:39:38,205][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 07:40:25,389][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-352 [2025-07-06 07:40:25,731][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 07:40:25,732][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 07:40:26,767][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-320] due to args.save_total_limit [2025-07-06 07:50:12,300][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 07:50:12,303][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 07:50:12,304][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 07:50:12,304][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 07:50:59,498][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-384 [2025-07-06 07:50:59,844][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 07:50:59,845][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 07:51:00,947][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-352] due to args.save_total_limit [2025-07-06 08:00:45,379][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 08:00:45,383][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 08:00:45,383][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 08:00:45,383][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 08:01:32,339][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-416 [2025-07-06 08:01:32,692][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 08:01:32,692][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 08:01:33,756][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-384] due to args.save_total_limit [2025-07-06 08:01:33,792][transformers.trainer][INFO] - Training completed. Do not forget to share your model on huggingface.co/models =) [2025-07-06 08:01:33,792][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-256 (score: 0.577060350421804). [2025-07-06 08:01:34,058][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-07-06/05-42-53/results/phi4-balanced/C2/checkpoint-416] due to args.save_total_limit [2025-07-06 08:01:34,095][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 08:01:34,103][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 08:01:34,103][transformers.trainer][INFO] - Num examples = 132 [2025-07-06 08:01:34,103][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 08:02:21,069][__main__][INFO] - Training completed successfully. [2025-07-06 08:02:21,069][__main__][INFO] - Running on Test [2025-07-06 08:02:21,069][transformers.trainer][INFO] - The following columns in the Evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference. If essay_year, id, grades, prompt, supporting_text, id_prompt, essay_text, reference are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message. [2025-07-06 08:02:21,072][transformers.trainer][INFO] - ***** Running Evaluation ***** [2025-07-06 08:02:21,072][transformers.trainer][INFO] - Num examples = 138 [2025-07-06 08:02:21,072][transformers.trainer][INFO] - Batch size = 4 [2025-07-06 08:03:10,772][__main__][INFO] - Test metrics: {'eval_loss': 1.3918403387069702, 'eval_model_preparation_time': 0.0091, 'eval_accuracy': 0.5144927536231884, 'eval_RMSE': 52.08688393694587, 'eval_QWK': 0.5905356055995131, 'eval_HDIV': 0.050724637681159424, 'eval_Macro_F1': 0.36549678320993584, 'eval_Micro_F1': 0.5144927536231884, 'eval_Weighted_F1': 0.45549927273540014, 'eval_TP_0': 0, 'eval_TN_0': 137, 'eval_FP_0': 0, 'eval_FN_0': 1, 'eval_TP_1': 29, 'eval_TN_1': 78, 'eval_FP_1': 25, 'eval_FN_1': 6, 'eval_TP_2': 4, 'eval_TN_2': 125, 'eval_FP_2': 8, 'eval_FN_2': 1, 'eval_TP_3': 26, 'eval_TN_3': 65, 'eval_FP_3': 22, 'eval_FN_3': 25, 'eval_TP_4': 0, 'eval_TN_4': 112, 'eval_FP_4': 0, 'eval_FN_4': 26, 'eval_TP_5': 12, 'eval_TN_5': 106, 'eval_FP_5': 12, 'eval_FN_5': 8, 'eval_runtime': 49.6901, 'eval_samples_per_second': 2.777, 'eval_steps_per_second': 0.704, 'epoch': 13.0} [2025-07-06 08:03:10,772][transformers.trainer][INFO] - Saving model checkpoint to ./results/phi4-balanced/C2/best_model [2025-07-06 08:03:11,094][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 08:03:11,095][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 08:03:11,716][transformers.tokenization_utils_base][INFO] - chat template saved in ./results/phi4-balanced/C2/best_model/chat_template.jinja [2025-07-06 08:03:11,717][transformers.tokenization_utils_base][INFO] - tokenizer config file saved in ./results/phi4-balanced/C2/best_model/tokenizer_config.json [2025-07-06 08:03:11,717][transformers.tokenization_utils_base][INFO] - Special tokens file saved in ./results/phi4-balanced/C2/best_model/special_tokens_map.json [2025-07-06 08:03:11,787][__main__][INFO] - Model and tokenizer saved to ./results/phi4-balanced/C2/best_model [2025-07-06 08:03:11,800][__main__][INFO] - Fine Tuning Finished. [2025-07-06 08:03:12,308][__main__][INFO] - Total emissions: 0.1076 kg CO2eq