Pushing fine-tuned model to Hugging Face Hub

README.md

@@ -0,0 +1,49 @@
+
+---
+language:
+- pt
+- en
+tags:
+- aes
+datasets:
+- kamel-usp/aes_enem_dataset
+base_model: microsoft/phi-4
+metrics:
+- accuracy
+- qwk
+library_name: peft
+model-index:
+  - name: phi4-balanced-C4
+    results:
+      - task:
+          type: text-classification
+          name: Automated Essay Score
+        dataset:
+          name: Automated Essay Score ENEM Dataset
+          type: kamel-usp/aes_enem_dataset
+          config: JBCS2025
+          split: test
+        metrics:
+          - name: Macro F1 (ignoring nan)
+            type: f1
+            value: 0.3296276013143483
+          - name: QWK
+            type: qwk
+            value: 0.579465541490858
+          - name: Weighted Macro F1
+            type: f1
+            value: 0.6761155293109196
+---
+# Model ID: phi4-balanced-C4
+## Results
+|                              |   test_data |
+|:-----------------------------|------------:|
+| eval_accuracy                |  0.702899   |
+| eval_RMSE                    | 24.554      |
+| eval_QWK                     |  0.579466   |
+| eval_Macro_F1                |  0.27469    |
+| eval_Macro_F1_(ignoring_nan) |  0.329628   |
+| eval_Weighted_F1             |  0.676116   |
+| eval_Micro_F1                |  0.702899   |
+| eval_HDIV                    |  0.00724638 |
+        

adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "microsoft/phi-4",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": [
+    "classifier",
+    "score"
+  ],
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_up_proj",
+    "o_proj",
+    "down_proj",
+    "qkv_proj"
+  ],
+  "task_type": "SEQ_CLS",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b7eade13249326942b2996e0f9ddbb92e800c4307dd6457f5a544762c299b89
+size 111515584

run_experiment.log

@@ -0,0 +1,1008 @@
+[2025-03-23 20:41:58,914][__main__][INFO] - cache_dir: /media/data/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
+post_training_results:
+  model_path: /workspace/jbcs2025/outputs/2025-03-23/18-55-03
+experiments:
+  model:
+    name: microsoft/phi-4
+    type: phi4_classification_lora
+    num_labels: 6
+    output_dir: ./results/phi4-balanced/C4
+    logging_dir: ./logs/phi4-balanced/C4
+    best_model_dir: ./results/phi4-balanced/C4/best_model
+    lora_r: 8
+    lora_dropout: 0.05
+    lora_alpha: 16
+    lora_target_modules: all-linear
+  dataset:
+    grade_index: 3
+  training_id: phi4-balanced-C4
+  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
+
+[2025-03-23 20:41:58,916][__main__][INFO] - Starting the Fine Tuning training process.
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file vocab.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/vocab.json
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file merges.txt from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/merges.txt
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file tokenizer.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/tokenizer.json
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file added_tokens.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/added_tokens.json
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file special_tokens_map.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/special_tokens_map.json
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file tokenizer_config.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/tokenizer_config.json
+[2025-03-23 20:42:03,790][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None
+[2025-03-23 20:42:03,944][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False
+[2025-03-23 20:42:04,348][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 20:42:04,349][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": 0,
+    "1": 40,
+    "2": 80,
+    "3": 120,
+    "4": 160,
+    "5": 200
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 17920,
+  "label2id": {
+    "0": 0,
+    "40": 1,
+    "80": 2,
+    "120": 3,
+    "160": 4,
+    "200": 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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 20:42:04,371][transformers.modeling_utils][INFO] - loading weights file model.safetensors from cache at /media/data/tmp/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/model.safetensors.index.json
+[2025-03-23 20:42:04,372][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object
+[2025-03-23 20:42:04,372][transformers.modeling_utils][INFO] - Instantiating Phi3ForSequenceClassification model under default dtype torch.bfloat16.
+[2025-03-23 20:42:25,935][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).
+[2025-03-23 20:42:25,936][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.
+[2025-03-23 20:42:28,026][__main__][INFO] - None
+[2025-03-23 20:42:28,029][transformers.training_args][INFO] - PyTorch: setting up devices
+[2025-03-23 20:42:28,068][__main__][INFO] - Total steps: 620. Number of warmup steps: 62
+[2025-03-23 20:42:28,077][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.
+[2025-03-23 20:42:28,100][transformers.trainer][INFO] - Using auto half precision backend
+[2025-03-23 20:42:28,101][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.
+[2025-03-23 20:42:28,121][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 20:42:28,133][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 20:42:28,134][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 20:42:28,134][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 20:42:54,626][transformers][INFO] - {'accuracy': 0.36363636363636365, 'RMSE': 36.680438185149114, 'QWK': 0.0, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.10666666666666666, 'Micro_F1': 0.36363636363636365, 'Weighted_F1': 0.19393939393939394, 'Macro_F1_(ignoring_nan)': np.float64(0.5333333333333333)}
+[2025-03-23 20:42:54,629][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 20:42:54,875][transformers.trainer][INFO] - The following columns in the training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] - ***** Running training *****
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Num examples = 500
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Num Epochs = 20
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Instantaneous batch size per device = 1
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 16
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Gradient Accumulation steps = 16
+[2025-03-23 20:42:54,906][transformers.trainer][INFO] -   Total optimization steps = 620
+[2025-03-23 20:42:54,908][transformers.trainer][INFO] -   Number of trainable parameters = 27,883,520
+[2025-03-23 20:49:54,343][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 20:49:54,346][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 20:49:54,346][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 20:49:54,346][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 20:50:20,840][transformers][INFO] - {'accuracy': 0.36363636363636365, 'RMSE': 36.680438185149114, 'QWK': 0.0, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.10666666666666666, 'Micro_F1': 0.36363636363636365, 'Weighted_F1': 0.19393939393939394, 'Macro_F1_(ignoring_nan)': np.float64(0.5333333333333333)}
+[2025-03-23 20:50:20,840][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 20:50:20,843][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-32
+[2025-03-23 20:50:21,157][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 20:50:21,158][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 20:57:21,498][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 20:57:21,501][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 20:57:21,501][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 20:57:21,502][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 20:57:47,876][transformers][INFO] - {'accuracy': 0.4621212121212121, 'RMSE': 33.93398225253192, 'QWK': 0.15453074433656944, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.18675097862089732, 'Micro_F1': 0.4621212121212121, 'Weighted_F1': 0.3838931318606115, 'Macro_F1_(ignoring_nan)': np.float64(0.4668774465522433)}
+[2025-03-23 20:57:47,876][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 20:57:47,879][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-64
+[2025-03-23 20:57:48,481][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 20:57:48,482][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 20:57:49,222][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-32] due to args.save_total_limit
+[2025-03-23 21:04:48,693][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:04:48,695][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:04:48,695][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:04:48,695][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:05:15,051][transformers][INFO] - {'accuracy': 0.5378787878787878, 'RMSE': 30.944720996896347, 'QWK': 0.3034998664173123, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.228490990990991, 'Micro_F1': 0.5378787878787878, 'Weighted_F1': 0.47856947856947857, 'Macro_F1_(ignoring_nan)': np.float64(0.5712274774774775)}
+[2025-03-23 21:05:15,051][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:05:15,054][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-96
+[2025-03-23 21:05:15,393][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:05:15,394][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:05:16,040][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-64] due to args.save_total_limit
+[2025-03-23 21:12:15,345][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:12:15,348][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:12:15,348][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:12:15,348][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:12:41,768][transformers][INFO] - {'accuracy': 0.5454545454545454, 'RMSE': 31.71845844395036, 'QWK': 0.3009188361408882, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.21155390793945011, 'Micro_F1': 0.5454545454545454, 'Weighted_F1': 0.46934591512904766, 'Macro_F1_(ignoring_nan)': np.float64(0.5288847698486253)}
+[2025-03-23 21:12:41,769][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:12:41,772][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-128
+[2025-03-23 21:12:42,252][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:12:42,252][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:19:42,312][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:19:42,315][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:19:42,315][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:19:42,315][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:20:08,584][transformers][INFO] - {'accuracy': 0.5606060606060606, 'RMSE': 28.91995221924885, 'QWK': 0.41987261146496824, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.28522175409256106, 'Micro_F1': 0.5606060606060606, 'Weighted_F1': 0.5337214163959735, 'Macro_F1_(ignoring_nan)': np.float64(0.47536959015426844)}
+[2025-03-23 21:20:08,584][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:20:08,588][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-160
+[2025-03-23 21:20:08,878][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:20:08,879][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:20:09,612][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-96] due to args.save_total_limit
+[2025-03-23 21:20:09,650][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-128] due to args.save_total_limit
+[2025-03-23 21:27:09,228][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:27:09,230][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:27:09,230][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:27:09,230][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:27:35,620][transformers][INFO] - {'accuracy': 0.5227272727272727, 'RMSE': 30.550504633038933, 'QWK': 0.3370727889381686, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.22424242424242422, 'Micro_F1': 0.5227272727272727, 'Weighted_F1': 0.4830119375573921, 'Macro_F1_(ignoring_nan)': np.float64(0.5606060606060606)}
+[2025-03-23 21:27:35,620][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:27:35,623][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-192
+[2025-03-23 21:27:35,955][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:27:35,955][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:34:35,458][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:34:35,460][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:34:35,460][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:34:35,460][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:35:01,816][transformers][INFO] - {'accuracy': 0.5075757575757576, 'RMSE': 34.289321599553055, 'QWK': 0.47082162340882794, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.30565610859728504, 'Micro_F1': 0.5075757575757576, 'Weighted_F1': 0.5056903880433292, 'Macro_F1_(ignoring_nan)': np.float64(0.5094268476621417)}
+[2025-03-23 21:35:01,816][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:35:01,819][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-224
+[2025-03-23 21:35:02,175][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:35:02,175][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:35:02,951][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-160] due to args.save_total_limit
+[2025-03-23 21:35:02,986][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-192] due to args.save_total_limit
+[2025-03-23 21:42:02,769][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:42:02,772][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:42:02,772][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:42:02,772][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:42:29,172][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.402048319327731, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.2524590163934426, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.5384997516145057, 'Macro_F1_(ignoring_nan)': np.float64(0.6311475409836065)}
+[2025-03-23 21:42:29,172][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:42:29,175][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-256
+[2025-03-23 21:42:29,463][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:42:29,464][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:49:29,409][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:49:29,411][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:49:29,411][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:49:29,411][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:49:55,727][transformers][INFO] - {'accuracy': 0.5984848484848485, 'RMSE': 28.4977404739606, 'QWK': 0.4918409561020455, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3331918158567775, 'Micro_F1': 0.5984848484848485, 'Weighted_F1': 0.5808866155157715, 'Macro_F1_(ignoring_nan)': np.float64(0.5553196930946291)}
+[2025-03-23 21:49:55,727][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:49:55,731][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-288
+[2025-03-23 21:49:56,023][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:49:56,023][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 21:49:56,763][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-224] due to args.save_total_limit
+[2025-03-23 21:49:56,810][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-256] due to args.save_total_limit
+[2025-03-23 21:56:56,055][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 21:56:56,057][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 21:56:56,057][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 21:56:56,057][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 21:57:22,544][transformers][INFO] - {'accuracy': 0.5984848484848485, 'RMSE': 28.4977404739606, 'QWK': 0.4536693847294292, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3165505532462971, 'Micro_F1': 0.5984848484848485, 'Weighted_F1': 0.5739240659034582, 'Macro_F1_(ignoring_nan)': np.float64(0.5275842554104951)}
+[2025-03-23 21:57:22,545][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 21:57:22,548][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-320
+[2025-03-23 21:57:22,867][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 21:57:22,868][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:04:23,489][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:04:23,491][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:04:23,491][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:04:23,491][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:04:49,840][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.44558071585098613, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.30430075187969924, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.5574436090225563, 'Macro_F1_(ignoring_nan)': np.float64(0.5071679197994987)}
+[2025-03-23 22:04:49,841][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:04:49,844][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-352
+[2025-03-23 22:04:50,184][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:04:50,185][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:04:50,944][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-320] due to args.save_total_limit
+[2025-03-23 22:11:50,060][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:11:50,062][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:11:50,062][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:11:50,062][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:12:16,556][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.40126216145148574, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.2525806451612903, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.5380254154447703, 'Macro_F1_(ignoring_nan)': np.float64(0.6314516129032258)}
+[2025-03-23 22:12:16,556][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:12:16,559][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-384
+[2025-03-23 22:12:16,980][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:12:16,980][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:12:17,746][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-352] due to args.save_total_limit
+[2025-03-23 22:19:16,915][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:19:16,918][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:19:16,918][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:19:16,918][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:19:43,392][transformers][INFO] - {'accuracy': 0.5681818181818182, 'RMSE': 29.336088024923512, 'QWK': 0.5011709601873537, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3423423989280405, 'Micro_F1': 0.5681818181818182, 'Weighted_F1': 0.5600188811042401, 'Macro_F1_(ignoring_nan)': np.float64(0.5705706648800676)}
+[2025-03-23 22:19:43,393][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:19:43,396][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-416
+[2025-03-23 22:19:43,732][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:19:43,733][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:19:44,597][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-288] due to args.save_total_limit
+[2025-03-23 22:19:44,648][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-384] due to args.save_total_limit
+[2025-03-23 22:26:45,026][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:26:45,029][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:26:45,029][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:26:45,029][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:27:11,429][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.47181628392484354, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3308645677161419, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.56964396589584, 'Macro_F1_(ignoring_nan)': np.float64(0.5514409461935699)}
+[2025-03-23 22:27:11,430][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:27:11,432][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-448
+[2025-03-23 22:34:21,711][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:34:21,714][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:34:21,714][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:34:21,714][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:34:48,074][transformers][INFO] - {'accuracy': 0.5757575757575758, 'RMSE': 29.128763250176767, 'QWK': 0.4815396700706991, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3335450732209876, 'Micro_F1': 0.5757575757575758, 'Weighted_F1': 0.5647085320791996, 'Macro_F1_(ignoring_nan)': np.float64(0.5559084553683126)}
+[2025-03-23 22:34:48,075][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:34:48,076][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-480
+[2025-03-23 22:34:48,464][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:34:48,465][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:34:49,169][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-448] due to args.save_total_limit
+[2025-03-23 22:41:49,150][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:41:49,152][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:41:49,152][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:41:49,152][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:42:15,502][transformers][INFO] - {'accuracy': 0.5984848484848485, 'RMSE': 28.4977404739606, 'QWK': 0.47344605858537747, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.31150793650793646, 'Micro_F1': 0.5984848484848485, 'Weighted_F1': 0.5776815776815778, 'Macro_F1_(ignoring_nan)': np.float64(0.5191798941798941)}
+[2025-03-23 22:42:15,503][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:42:15,506][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-512
+[2025-03-23 22:42:16,316][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:42:16,317][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:42:17,055][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-480] due to args.save_total_limit
+[2025-03-23 22:49:17,118][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:49:17,120][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:49:17,120][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:49:17,120][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:49:43,519][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.48693105002253256, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3384561403508772, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.5721424774056353, 'Macro_F1_(ignoring_nan)': np.float64(0.564093567251462)}
+[2025-03-23 22:49:43,520][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:49:43,523][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-544
+[2025-03-23 22:49:43,837][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:49:43,837][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:49:44,524][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-512] due to args.save_total_limit
+[2025-03-23 22:56:43,629][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 22:56:43,631][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 22:56:43,631][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 22:56:43,631][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 22:57:09,913][transformers][INFO] - {'accuracy': 0.5984848484848485, 'RMSE': 28.4977404739606, 'QWK': 0.5040376850605652, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3435825471698113, 'Micro_F1': 0.5984848484848485, 'Weighted_F1': 0.5861349342481418, 'Macro_F1_(ignoring_nan)': np.float64(0.5726375786163521)}
+[2025-03-23 22:57:09,914][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 22:57:09,916][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-576
+[2025-03-23 22:57:11,876][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 22:57:11,877][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 22:57:12,612][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-416] due to args.save_total_limit
+[2025-03-23 22:57:12,654][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-544] due to args.save_total_limit
+[2025-03-23 23:04:12,951][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 23:04:12,954][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 23:04:12,954][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 23:04:12,954][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 23:04:39,310][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.47181628392484354, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3308645677161419, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.56964396589584, 'Macro_F1_(ignoring_nan)': np.float64(0.5514409461935699)}
+[2025-03-23 23:04:39,311][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 23:04:39,313][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-608
+[2025-03-23 23:04:39,620][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 23:04:39,621][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 23:07:22,340][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 23:07:22,342][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 23:07:22,342][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 23:07:22,342][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 23:07:48,658][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 28.91995221924885, 'QWK': 0.47181628392484354, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3308645677161419, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.56964396589584, 'Macro_F1_(ignoring_nan)': np.float64(0.5514409461935699)}
+[2025-03-23 23:07:48,659][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 23:07:48,661][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-620
+[2025-03-23 23:07:49,135][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 23:07:49,136][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 23:07:49,714][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-608] due to args.save_total_limit
+[2025-03-23 23:07:49,754][transformers.trainer][INFO] - 
+
+Training completed. Do not forget to share your model on huggingface.co/models =)
+
+
+[2025-03-23 23:07:49,754][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-576 (score: 0.5040376850605652).
+[2025-03-23 23:07:49,908][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/20-41-58/results/phi4-balanced/C4/checkpoint-620] due to args.save_total_limit
+[2025-03-23 23:07:49,954][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 23:07:49,957][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 23:07:49,957][transformers.trainer][INFO] -   Num examples = 132
+[2025-03-23 23:07:49,957][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 23:08:16,365][transformers][INFO] - {'accuracy': 0.5984848484848485, 'RMSE': 28.4977404739606, 'QWK': 0.5040376850605652, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.3435825471698113, 'Micro_F1': 0.5984848484848485, 'Weighted_F1': 0.5861349342481418, 'Macro_F1_(ignoring_nan)': np.float64(0.5726375786163521)}
+[2025-03-23 23:08:16,368][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 23:08:16,369][__main__][INFO] - Training completed successfully.
+[2025-03-23 23:08:16,369][__main__][INFO] - Running on Test
+[2025-03-23 23:08:16,369][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id. If essay_text, prompt, id_prompt, reference, grades, supporting_text, essay_year, id are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
+[2025-03-23 23:08:16,371][transformers.trainer][INFO] - 
+***** Running Evaluation *****
+[2025-03-23 23:08:16,371][transformers.trainer][INFO] -   Num examples = 138
+[2025-03-23 23:08:16,371][transformers.trainer][INFO] -   Batch size = 16
+[2025-03-23 23:08:44,397][transformers][INFO] - {'accuracy': 0.7028985507246377, 'RMSE': 24.55399256179405, 'QWK': 0.579465541490858, 'HDIV': 0.007246376811594235, 'Macro_F1': 0.27468966776195697, 'Micro_F1': 0.7028985507246377, 'Weighted_F1': 0.6761155293109196, 'Macro_F1_(ignoring_nan)': np.float64(0.3296276013143483)}
+[2025-03-23 23:08:44,398][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
+[2025-03-23 23:08:44,400][transformers.trainer][INFO] - Saving model checkpoint to ./results/phi4-balanced/C4/best_model
+[2025-03-23 23:08:44,680][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--microsoft--phi-4/snapshots/187ef0342fff0eb3333be9f00389385e95ef0b61/config.json
+[2025-03-23 23:08:44,681][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": "4.50.0",
+  "use_cache": true,
+  "vocab_size": 100352
+}
+
+[2025-03-23 23:08:45,141][__main__][INFO] - Fine Tuning Finished.

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbfdfe3dca0425b9fe06367acde7660a01e10437ac0937a022b0b6e8902fb8e3
+size 5432