[2025-03-31 21:52:55,059][__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/2025-03-31/19-27-08
experiments:
  model:
    name: meta-llama/Llama-3.1-8B
    type: llama31_classification_lora
    use_essay_prompt: true
    num_labels: 6
    output_dir: ./results/llama31-8b-balanced/C2
    logging_dir: ./logs/llama31-8b-balanced/C2
    best_model_dir: ./results/llama31-8b-balanced/C2/best_model
    lora_r: 8
    lora_dropout: 0.05
    lora_alpha: 16
    lora_target_modules: all-linear
  dataset:
    grade_index: 1
  training_id: llama31-8b-essay-prompt-balanced-C2
  training_params:
    weight_decay: 0.01
    warmup_ratio: 0.1
    learning_rate: 5.0e-05
    train_batch_size: 1
    eval_batch_size: 4
    gradient_accumulation_steps: 16
    gradient_checkpointing: false

[2025-03-31 21:52:55,062][__main__][INFO] - Starting the Fine Tuning training process.
[2025-03-31 21:53:00,354][transformers.tokenization_utils_base][INFO] - loading file tokenizer.json from cache at /media/data/tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/tokenizer.json
[2025-03-31 21:53:00,354][transformers.tokenization_utils_base][INFO] - loading file tokenizer.model from cache at None
[2025-03-31 21:53:00,354][transformers.tokenization_utils_base][INFO] - loading file added_tokens.json from cache at None
[2025-03-31 21:53:00,355][transformers.tokenization_utils_base][INFO] - loading file special_tokens_map.json from cache at /media/data/tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/special_tokens_map.json
[2025-03-31 21:53:00,355][transformers.tokenization_utils_base][INFO] - loading file tokenizer_config.json from cache at /media/data/tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/tokenizer_config.json
[2025-03-31 21:53:00,355][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None
[2025-03-31 21:53:00,740][transformers.tokenization_utils_base][INFO] - Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
[2025-03-31 21:53:00,748][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False
[2025-03-31 21:53:01,894][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /media/data/tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 21:53:01,896][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "id2label": {
    "0": 0,
    "1": 40,
    "2": 80,
    "3": 120,
    "4": 160,
    "5": 200
  },
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "label2id": {
    "0": 0,
    "40": 1,
    "80": 2,
    "120": 3,
    "160": 4,
    "200": 5
  },
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 21:53:01,924][transformers.modeling_utils][INFO] - loading weights file model.safetensors from cache at /media/data/tmp/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/model.safetensors.index.json
[2025-03-31 21:53:01,924][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object
[2025-03-31 21:53:01,924][transformers.modeling_utils][INFO] - Instantiating LlamaForSequenceClassification model under default dtype torch.bfloat16.
[2025-03-31 21:53:08,644][transformers.modeling_utils][INFO] - Some weights of the model checkpoint at meta-llama/Llama-3.1-8B were not used when initializing LlamaForSequenceClassification: ['lm_head.weight']
- This IS expected if you are initializing LlamaForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing LlamaForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
[2025-03-31 21:53:08,644][transformers.modeling_utils][WARNING] - Some weights of LlamaForSequenceClassification were not initialized from the model checkpoint at meta-llama/Llama-3.1-8B and are newly initialized: ['score.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
[2025-03-31 21:53:09,692][__main__][INFO] - None
[2025-03-31 21:53:09,693][transformers.training_args][INFO] - PyTorch: setting up devices
[2025-03-31 21:53:09,711][__main__][INFO] - Total steps: 620. Number of warmup steps: 62
[2025-03-31 21:53:09,717][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.
[2025-03-31 21:53:09,736][transformers.trainer][INFO] - Using auto half precision backend
[2025-03-31 21:53:09,737][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.
[2025-03-31 21:53:09,763][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 21:53:09,780][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 21:53:09,780][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 21:53:09,780][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 21:53:40,822][transformers][INFO] - {'accuracy': 0.42424242424242425, 'RMSE': 57.735026918962575, 'QWK': -0.1321107784431137, 'HDIV': 0.045454545454545414, 'Macro_F1': 0.13404696132596686, 'Micro_F1': 0.42424242424242425, 'Weighted_F1': 0.29505169094257494, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(97), 'FP_1': np.int64(10), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(55), 'TN_3': np.int64(6), 'FP_3': np.int64(65), 'FN_3': np.int64(6), 'TP_4': np.int64(1), 'TN_4': np.int64(101), 'FP_4': np.int64(1), 'FN_4': np.int64(29), 'TP_5': np.int64(0), 'TN_5': np.int64(117), 'FP_5': np.int64(0), 'FN_5': np.int64(15)}
[2025-03-31 21:53:41,037][transformers.trainer][INFO] - The following columns in the training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 21:53:41,083][transformers.trainer][INFO] - ***** Running training *****
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Num examples = 500
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Num Epochs = 20
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Instantaneous batch size per device = 1
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 16
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Gradient Accumulation steps = 16
[2025-03-31 21:53:41,083][transformers.trainer][INFO] -   Total optimization steps = 620
[2025-03-31 21:53:41,086][transformers.trainer][INFO] -   Number of trainable parameters = 20,996,096
[2025-03-31 22:01:23,841][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:01:23,843][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:01:23,844][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:01:23,844][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:01:54,478][transformers][INFO] - {'accuracy': 0.4621212121212121, 'RMSE': 49.11335065052284, 'QWK': 0.0, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.12642487046632125, 'Micro_F1': 0.4621212121212121, 'Weighted_F1': 0.29211807191081807, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(107), 'FP_1': np.int64(0), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(61), 'TN_3': np.int64(0), 'FP_3': np.int64(71), 'FN_3': np.int64(0), 'TP_4': np.int64(0), 'TN_4': np.int64(102), 'FP_4': np.int64(0), 'FN_4': np.int64(30), 'TP_5': np.int64(0), 'TN_5': np.int64(117), 'FP_5': np.int64(0), 'FN_5': np.int64(15)}
[2025-03-31 22:01:54,484][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-32
[2025-03-31 22:01:55,051][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:01:55,052][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:09:38,048][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:09:38,050][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:09:38,050][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:09:38,050][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:10:08,712][transformers][INFO] - {'accuracy': 0.44696969696969696, 'RMSE': 50.81159495448044, 'QWK': -0.009478672985782088, 'HDIV': 0.015151515151515138, 'Macro_F1': 0.1358288770053476, 'Micro_F1': 0.44696969696969696, 'Weighted_F1': 0.30003240965807815, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(105), 'FP_1': np.int64(2), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(58), 'TN_3': np.int64(3), 'FP_3': np.int64(68), 'FN_3': np.int64(3), 'TP_4': np.int64(1), 'TN_4': np.int64(99), 'FP_4': np.int64(3), 'FN_4': np.int64(29), 'TP_5': np.int64(0), 'TN_5': np.int64(117), 'FP_5': np.int64(0), 'FN_5': np.int64(15)}
[2025-03-31 22:10:08,717][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-64
[2025-03-31 22:10:09,113][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:10:09,114][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:17:51,028][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:17:51,030][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:17:51,030][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:17:51,030][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:18:21,719][transformers][INFO] - {'accuracy': 0.42424242424242425, 'RMSE': 56.889046716952784, 'QWK': 0.06544336020364871, 'HDIV': 0.030303030303030276, 'Macro_F1': 0.19495079161317927, 'Micro_F1': 0.42424242424242425, 'Weighted_F1': 0.34293188625666166, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(107), 'FP_1': np.int64(0), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(49), 'TN_3': np.int64(17), 'FP_3': np.int64(54), 'FN_3': np.int64(12), 'TP_4': np.int64(4), 'TN_4': np.int64(98), 'FP_4': np.int64(4), 'FN_4': np.int64(26), 'TP_5': np.int64(3), 'TN_5': np.int64(99), 'FP_5': np.int64(18), 'FN_5': np.int64(12)}
[2025-03-31 22:18:21,726][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-96
[2025-03-31 22:18:22,286][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:18:22,287][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:18:22,519][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-32] due to args.save_total_limit
[2025-03-31 22:18:22,528][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-64] due to args.save_total_limit
[2025-03-31 22:26:04,029][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:26:04,032][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:26:04,032][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:26:04,032][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:26:34,632][transformers][INFO] - {'accuracy': 0.45454545454545453, 'RMSE': 51.99067515459193, 'QWK': 0.024005305039787816, 'HDIV': 0.022727272727272707, 'Macro_F1': 0.1497311827956989, 'Micro_F1': 0.45454545454545453, 'Weighted_F1': 0.3118788693385468, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(102), 'FP_1': np.int64(5), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(58), 'TN_3': np.int64(4), 'FP_3': np.int64(67), 'FN_3': np.int64(3), 'TP_4': np.int64(0), 'TN_4': np.int64(102), 'FP_4': np.int64(0), 'FN_4': np.int64(30), 'TP_5': np.int64(0), 'TN_5': np.int64(117), 'FP_5': np.int64(0), 'FN_5': np.int64(15)}
[2025-03-31 22:26:34,639][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-128
[2025-03-31 22:26:35,015][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:26:35,016][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:34:16,743][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:34:16,745][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:34:16,745][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:34:16,745][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:34:47,324][transformers][INFO] - {'accuracy': 0.4621212121212121, 'RMSE': 50.81159495448044, 'QWK': 0.32080394240989474, 'HDIV': 0.022727272727272707, 'Macro_F1': 0.1899047698896784, 'Micro_F1': 0.4621212121212121, 'Weighted_F1': 0.36297436415338696, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(103), 'FP_1': np.int64(4), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(130), 'FP_2': np.int64(2), 'FN_2': np.int64(0), 'TP_3': np.int64(52), 'TN_3': np.int64(22), 'FP_3': np.int64(49), 'FN_3': np.int64(9), 'TP_4': np.int64(0), 'TN_4': np.int64(102), 'FP_4': np.int64(0), 'FN_4': np.int64(30), 'TP_5': np.int64(7), 'TN_5': np.int64(101), 'FP_5': np.int64(16), 'FN_5': np.int64(8)}
[2025-03-31 22:34:47,331][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-160
[2025-03-31 22:34:47,788][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:34:47,790][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:34:48,021][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-96] due to args.save_total_limit
[2025-03-31 22:34:48,023][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-128] due to args.save_total_limit
[2025-03-31 22:42:29,624][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:42:29,626][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:42:29,626][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:42:29,626][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:43:00,194][transformers][INFO] - {'accuracy': 0.3106060606060606, 'RMSE': 62.667956144051224, 'QWK': 0.1684877707353113, 'HDIV': 0.14393939393939392, 'Macro_F1': 0.15488203701629263, 'Micro_F1': 0.3106060606060606, 'Weighted_F1': 0.26871220748993235, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(5), 'TN_1': np.int64(89), 'FP_1': np.int64(18), 'FN_1': np.int64(20), 'TP_2': np.int64(0), 'TN_2': np.int64(128), 'FP_2': np.int64(4), 'FN_2': np.int64(0), 'TP_3': np.int64(11), 'TN_3': np.int64(64), 'FP_3': np.int64(7), 'FN_3': np.int64(50), 'TP_4': np.int64(25), 'TN_4': np.int64(44), 'FP_4': np.int64(58), 'FN_4': np.int64(5), 'TP_5': np.int64(0), 'TN_5': np.int64(113), 'FP_5': np.int64(4), 'FN_5': np.int64(15)}
[2025-03-31 22:43:00,200][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-192
[2025-03-31 22:43:01,040][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:43:01,041][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:50:42,643][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:50:42,645][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:50:42,645][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:50:42,645][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:51:13,207][transformers][INFO] - {'accuracy': 0.3484848484848485, 'RMSE': 60.10092521851234, 'QWK': 0.21217704786701397, 'HDIV': 0.08333333333333337, 'Macro_F1': 0.18135976295114997, 'Micro_F1': 0.3484848484848485, 'Weighted_F1': 0.33908939082082523, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(101), 'FP_1': np.int64(6), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(130), 'FP_2': np.int64(2), 'FN_2': np.int64(0), 'TP_3': np.int64(33), 'TN_3': np.int64(42), 'FP_3': np.int64(29), 'FN_3': np.int64(28), 'TP_4': np.int64(5), 'TN_4': np.int64(78), 'FP_4': np.int64(24), 'FN_4': np.int64(25), 'TP_5': np.int64(6), 'TN_5': np.int64(92), 'FP_5': np.int64(25), 'FN_5': np.int64(9)}
[2025-03-31 22:51:13,214][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-224
[2025-03-31 22:51:13,569][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:51:13,570][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:51:13,818][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-192] due to args.save_total_limit
[2025-03-31 22:58:55,643][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 22:58:55,645][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 22:58:55,645][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 22:58:55,645][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 22:59:26,224][transformers][INFO] - {'accuracy': 0.4393939393939394, 'RMSE': 51.99067515459193, 'QWK': 0.333846293111252, 'HDIV': 0.037878787878787845, 'Macro_F1': 0.2625054945054945, 'Micro_F1': 0.4393939393939394, 'Weighted_F1': 0.3995337995337996, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(99), 'FP_1': np.int64(8), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(44), 'TN_3': np.int64(33), 'FP_3': np.int64(38), 'FN_3': np.int64(17), 'TP_4': np.int64(6), 'TN_4': np.int64(88), 'FP_4': np.int64(14), 'FN_4': np.int64(24), 'TP_5': np.int64(6), 'TN_5': np.int64(103), 'FP_5': np.int64(14), 'FN_5': np.int64(9)}
[2025-03-31 22:59:26,231][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-256
[2025-03-31 22:59:26,649][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 22:59:26,650][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 22:59:26,941][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-160] due to args.save_total_limit
[2025-03-31 22:59:26,947][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-224] due to args.save_total_limit
[2025-03-31 23:07:08,652][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:07:08,654][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:07:08,654][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:07:08,654][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:07:39,231][transformers][INFO] - {'accuracy': 0.4621212121212121, 'RMSE': 52.223296786709355, 'QWK': 0.2920480549199085, 'HDIV': 0.045454545454545414, 'Macro_F1': 0.20456615634883887, 'Micro_F1': 0.4621212121212121, 'Weighted_F1': 0.38994689692328216, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(101), 'FP_1': np.int64(6), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(130), 'FP_2': np.int64(2), 'FN_2': np.int64(0), 'TP_3': np.int64(51), 'TN_3': np.int64(28), 'FP_3': np.int64(43), 'FN_3': np.int64(10), 'TP_4': np.int64(2), 'TN_4': np.int64(96), 'FP_4': np.int64(6), 'FN_4': np.int64(28), 'TP_5': np.int64(6), 'TN_5': np.int64(103), 'FP_5': np.int64(14), 'FN_5': np.int64(9)}
[2025-03-31 23:07:39,238][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-288
[2025-03-31 23:07:39,662][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:07:39,662][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:15:21,392][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:15:21,394][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:15:21,394][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:15:21,395][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:15:51,966][transformers][INFO] - {'accuracy': 0.32575757575757575, 'RMSE': 58.15340215170718, 'QWK': 0.23574333859052043, 'HDIV': 0.06818181818181823, 'Macro_F1': 0.20009987814865865, 'Micro_F1': 0.32575757575757575, 'Weighted_F1': 0.3112339309678556, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(95), 'FP_1': np.int64(12), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(30), 'TN_3': np.int64(39), 'FP_3': np.int64(32), 'FN_3': np.int64(31), 'TP_4': np.int64(8), 'TN_4': np.int64(74), 'FP_4': np.int64(28), 'FN_4': np.int64(22), 'TP_5': np.int64(5), 'TN_5': np.int64(100), 'FP_5': np.int64(17), 'FN_5': np.int64(10)}
[2025-03-31 23:15:51,973][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-320
[2025-03-31 23:15:52,378][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:15:52,379][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:15:52,621][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-288] due to args.save_total_limit
[2025-03-31 23:23:34,424][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:23:34,426][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:23:34,426][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:23:34,426][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:24:05,005][transformers][INFO] - {'accuracy': 0.3181818181818182, 'RMSE': 59.49280577846316, 'QWK': 0.21788888437969256, 'HDIV': 0.05303030303030298, 'Macro_F1': 0.17886917100177863, 'Micro_F1': 0.3181818181818182, 'Weighted_F1': 0.2925910312723771, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(91), 'FP_1': np.int64(16), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(34), 'TN_3': np.int64(30), 'FP_3': np.int64(41), 'FN_3': np.int64(27), 'TP_4': np.int64(2), 'TN_4': np.int64(87), 'FP_4': np.int64(15), 'FN_4': np.int64(28), 'TP_5': np.int64(4), 'TN_5': np.int64(99), 'FP_5': np.int64(18), 'FN_5': np.int64(11)}
[2025-03-31 23:24:05,012][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-352
[2025-03-31 23:24:05,359][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:24:05,359][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:24:05,625][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-320] due to args.save_total_limit
[2025-03-31 23:31:47,429][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:31:47,431][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:31:47,431][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:31:47,431][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:32:17,996][transformers][INFO] - {'accuracy': 0.3787878787878788, 'RMSE': 58.77538136452586, 'QWK': 0.18103448275862066, 'HDIV': 0.0757575757575758, 'Macro_F1': 0.2044195455960162, 'Micro_F1': 0.3787878787878788, 'Weighted_F1': 0.3379735732676909, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(101), 'FP_1': np.int64(6), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(40), 'TN_3': np.int64(36), 'FP_3': np.int64(35), 'FN_3': np.int64(21), 'TP_4': np.int64(4), 'TN_4': np.int64(82), 'FP_4': np.int64(20), 'FN_4': np.int64(26), 'TP_5': np.int64(6), 'TN_5': np.int64(96), 'FP_5': np.int64(21), 'FN_5': np.int64(9)}
[2025-03-31 23:32:18,003][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-384
[2025-03-31 23:32:18,391][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:32:18,392][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:32:18,621][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-352] due to args.save_total_limit
[2025-03-31 23:40:00,223][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:40:00,225][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:40:00,225][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:40:00,225][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:40:30,805][transformers][INFO] - {'accuracy': 0.2803030303030303, 'RMSE': 64.9475312708791, 'QWK': 0.18376630299584207, 'HDIV': 0.08333333333333337, 'Macro_F1': 0.16438326291399755, 'Micro_F1': 0.2803030303030303, 'Weighted_F1': 0.27356487045733335, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(0), 'TN_1': np.int64(97), 'FP_1': np.int64(10), 'FN_1': np.int64(25), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(27), 'TN_3': np.int64(44), 'FP_3': np.int64(27), 'FN_3': np.int64(34), 'TP_4': np.int64(4), 'TN_4': np.int64(81), 'FP_4': np.int64(21), 'FN_4': np.int64(26), 'TP_5': np.int64(6), 'TN_5': np.int64(80), 'FP_5': np.int64(37), 'FN_5': np.int64(9)}
[2025-03-31 23:40:30,811][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-416
[2025-03-31 23:40:31,234][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:40:31,235][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:40:31,523][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-384] due to args.save_total_limit
[2025-03-31 23:40:31,529][transformers.trainer][INFO] - 

Training completed. Do not forget to share your model on huggingface.co/models =)


[2025-03-31 23:40:31,530][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-256 (score: 0.333846293111252).
[2025-03-31 23:40:31,609][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-31/21-52-55/results/llama31-8b-balanced/C2/checkpoint-416] due to args.save_total_limit
[2025-03-31 23:40:31,616][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:40:31,618][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:40:31,619][transformers.trainer][INFO] -   Num examples = 132
[2025-03-31 23:40:31,619][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:41:02,204][transformers][INFO] - {'accuracy': 0.4393939393939394, 'RMSE': 51.99067515459193, 'QWK': 0.333846293111252, 'HDIV': 0.037878787878787845, 'Macro_F1': 0.2625054945054945, 'Micro_F1': 0.4393939393939394, 'Weighted_F1': 0.3995337995337996, 'TP_0': np.int64(0), 'TN_0': np.int64(131), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(2), 'TN_1': np.int64(99), 'FP_1': np.int64(8), 'FN_1': np.int64(23), 'TP_2': np.int64(0), 'TN_2': np.int64(132), 'FP_2': np.int64(0), 'FN_2': np.int64(0), 'TP_3': np.int64(44), 'TN_3': np.int64(33), 'FP_3': np.int64(38), 'FN_3': np.int64(17), 'TP_4': np.int64(6), 'TN_4': np.int64(88), 'FP_4': np.int64(14), 'FN_4': np.int64(24), 'TP_5': np.int64(6), 'TN_5': np.int64(103), 'FP_5': np.int64(14), 'FN_5': np.int64(9)}
[2025-03-31 23:41:02,210][__main__][INFO] - Training completed successfully.
[2025-03-31 23:41:02,210][__main__][INFO] - Running on Test
[2025-03-31 23:41:02,210][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text. If prompt, supporting_text, id, id_prompt, essay_year, reference, grades, essay_text are not expected by `PeftModelForSequenceClassification.forward`,  you can safely ignore this message.
[2025-03-31 23:41:02,212][transformers.trainer][INFO] - 
***** Running Evaluation *****
[2025-03-31 23:41:02,212][transformers.trainer][INFO] -   Num examples = 138
[2025-03-31 23:41:02,212][transformers.trainer][INFO] -   Batch size = 4
[2025-03-31 23:41:34,456][transformers][INFO] - {'accuracy': 0.38405797101449274, 'RMSE': 56.568542494923804, 'QWK': 0.34285714285714297, 'HDIV': 0.02898550724637683, 'Macro_F1': 0.2263684831084438, 'Micro_F1': 0.38405797101449274, 'Weighted_F1': 0.35160147897762273, 'TP_0': np.int64(0), 'TN_0': np.int64(137), 'FP_0': np.int64(0), 'FN_0': np.int64(1), 'TP_1': np.int64(7), 'TN_1': np.int64(88), 'FP_1': np.int64(15), 'FN_1': np.int64(28), 'TP_2': np.int64(0), 'TN_2': np.int64(133), 'FP_2': np.int64(0), 'FN_2': np.int64(5), 'TP_3': np.int64(34), 'TN_3': np.int64(35), 'FP_3': np.int64(52), 'FN_3': np.int64(17), 'TP_4': np.int64(8), 'TN_4': np.int64(104), 'FP_4': np.int64(8), 'FN_4': np.int64(18), 'TP_5': np.int64(4), 'TN_5': np.int64(108), 'FP_5': np.int64(10), 'FN_5': np.int64(16)}
[2025-03-31 23:41:34,461][transformers.trainer][INFO] - Saving model checkpoint to ./results/llama31-8b-balanced/C2/best_model
[2025-03-31 23:41:34,820][transformers.configuration_utils][INFO] - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--meta-llama--Llama-3.1-8B/snapshots/d04e592bb4f6aa9cfee91e2e20afa771667e1d4b/config.json
[2025-03-31 23:41:34,821][transformers.configuration_utils][INFO] - Model config LlamaConfig {
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.50.3",
  "use_cache": true,
  "vocab_size": 128256
}

[2025-03-31 23:41:34,905][__main__][INFO] - Fine Tuning Finished.