jbcs2025_phi4-balanced-C1 / run_experiment.log
abarbosa's picture
Pushing fine-tuned model to Hugging Face Hub
877cacf verified
[2025-03-23 15:04:12,688][__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/2025-03-17/22-47-57
experiments:
model:
name: microsoft/phi-4
type: phi4_classification_lora
num_labels: 6
output_dir: ./results/phi4-balanced/C1
logging_dir: ./logs/phi4-balanced/C1
best_model_dir: ./results/phi4-balanced/C1/best_model
lora_r: 8
lora_dropout: 0.05
lora_alpha: 16
lora_target_modules: all-linear
dataset:
grade_index: 0
training_id: phi4-balanced-C1
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 15:04:12,690][__main__][INFO] - Starting the Fine Tuning training process.
[2025-03-23 15:04:15,506][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 15:04:15,506][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 15:04:15,506][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 15:04:15,506][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 15:04:15,506][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 15:04:15,506][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 15:04:15,506][transformers.tokenization_utils_base][INFO] - loading file chat_template.jinja from cache at None
[2025-03-23 15:04:15,635][__main__][INFO] - Tokenizer function parameters- Padding:longest; Truncation: False
[2025-03-23 15:04:15,839][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 15:04:15,839][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 15:04:15,860][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 15:04:15,861][transformers.modeling_utils][INFO] - Will use torch_dtype=torch.bfloat16 as defined in model's config object
[2025-03-23 15:04:15,861][transformers.modeling_utils][INFO] - Instantiating Phi3ForSequenceClassification model under default dtype torch.bfloat16.
[2025-03-23 15:04:37,315][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 15:04:37,316][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 15:04:42,122][__main__][INFO] - None
[2025-03-23 15:04:42,123][transformers.training_args][INFO] - PyTorch: setting up devices
[2025-03-23 15:04:42,177][__main__][INFO] - Total steps: 620. Number of warmup steps: 62
[2025-03-23 15:04:42,183][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 15:04:42,201][transformers.trainer][INFO] - Using auto half precision backend
[2025-03-23 15:04:42,202][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 15:04:42,218][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:04:42,230][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:04:42,230][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:04:42,230][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:05:05,414][transformers][INFO] - {'accuracy': 0.4621212121212121, 'RMSE': 44.17595671307132, 'QWK': -0.25277057297807115, 'HDIV': 0.05303030303030298, 'Macro_F1': 0.1304812834224599, 'Micro_F1': 0.4621212121212121, 'Weighted_F1': 0.31137578998541565, 'Macro_F1_(ignoring_nan)': np.float64(0.32620320855614976)}
[2025-03-23 15:05:05,416][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:05:05,636][transformers.trainer][INFO] - The following columns in the training set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:05:05,666][transformers.trainer][INFO] - ***** Running training *****
[2025-03-23 15:05:05,666][transformers.trainer][INFO] - Num examples = 500
[2025-03-23 15:05:05,667][transformers.trainer][INFO] - Num Epochs = 20
[2025-03-23 15:05:05,667][transformers.trainer][INFO] - Instantaneous batch size per device = 1
[2025-03-23 15:05:05,667][transformers.trainer][INFO] - Total train batch size (w. parallel, distributed & accumulation) = 16
[2025-03-23 15:05:05,667][transformers.trainer][INFO] - Gradient Accumulation steps = 16
[2025-03-23 15:05:05,667][transformers.trainer][INFO] - Total optimization steps = 620
[2025-03-23 15:05:05,669][transformers.trainer][INFO] - Number of trainable parameters = 27,883,520
[2025-03-23 15:11:20,753][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:11:20,755][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:11:20,755][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:11:20,755][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:11:43,814][transformers][INFO] - {'accuracy': 0.4772727272727273, 'RMSE': 35.67530340063379, 'QWK': 0.0, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.12923076923076923, 'Micro_F1': 0.4772727272727273, 'Weighted_F1': 0.30839160839160845, 'Macro_F1_(ignoring_nan)': np.float64(0.6461538461538462)}
[2025-03-23 15:11:43,814][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:11:43,816][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-32
[2025-03-23 15:11:44,193][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 15:11:44,194][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 15:18:00,654][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:18:00,657][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:18:00,657][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:18:00,657][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:18:23,632][transformers][INFO] - {'accuracy': 0.49242424242424243, 'RMSE': 35.33390508399614, 'QWK': 0.03629146583498721, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.15774990868135882, 'Micro_F1': 0.49242424242424243, 'Weighted_F1': 0.351078666858528, 'Macro_F1_(ignoring_nan)': np.float64(0.3943747717033971)}
[2025-03-23 15:18:23,632][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:18:23,635][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-64
[2025-03-23 15:18:23,868][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 15:18:23,869][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 15:18:24,533][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-32] due to args.save_total_limit
[2025-03-23 15:24:39,278][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:24:39,280][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:24:39,280][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:24:39,280][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:25:02,286][transformers][INFO] - {'accuracy': 0.49242424242424243, 'RMSE': 31.333978072025612, 'QWK': 0.4662539936102237, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.2550461466497952, 'Micro_F1': 0.49242424242424243, 'Weighted_F1': 0.44255470368357247, 'Macro_F1_(ignoring_nan)': np.float64(0.425076911082992)}
[2025-03-23 15:25:02,286][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:25:02,289][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-96
[2025-03-23 15:25:02,557][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 15:25:02,558][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 15:25:03,210][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-64] due to args.save_total_limit
[2025-03-23 15:31:17,930][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:31:17,932][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:31:17,932][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:31:17,933][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:31:40,845][transformers][INFO] - {'accuracy': 0.6590909090909091, 'RMSE': 26.74231693686086, 'QWK': 0.6705583756345177, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.505355276907001, 'Micro_F1': 0.6590909090909091, 'Weighted_F1': 0.655035860169089, 'Macro_F1_(ignoring_nan)': np.float64(0.6316940961337513)}
[2025-03-23 15:31:40,845][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:31:40,847][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-128
[2025-03-23 15:31:41,099][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 15:31:41,100][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 15:31:41,822][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-96] due to args.save_total_limit
[2025-03-23 15:37:56,610][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:37:56,613][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:37:56,613][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:37:56,613][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:38:19,586][transformers][INFO] - {'accuracy': 0.6742424242424242, 'RMSE': 25.81988897471611, 'QWK': 0.6984548928393421, 'HDIV': 0.0, 'Macro_F1': 0.5295884500497843, 'Micro_F1': 0.6742424242424242, 'Weighted_F1': 0.6732932544830984, 'Macro_F1_(ignoring_nan)': np.float64(0.6619855625622303)}
[2025-03-23 15:38:19,586][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:38:19,588][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-160
[2025-03-23 15:38:19,821][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 15:38:19,822][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 15:38:20,534][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-128] due to args.save_total_limit
[2025-03-23 15:44:35,255][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:44:35,256][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:44:35,256][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:44:35,256][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:44:58,245][transformers][INFO] - {'accuracy': 0.5833333333333334, 'RMSE': 27.633971188310298, 'QWK': 0.5477485316510767, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.30369105025593573, 'Micro_F1': 0.5833333333333334, 'Weighted_F1': 0.5354999307115893, 'Macro_F1_(ignoring_nan)': np.float64(0.5061517504265596)}
[2025-03-23 15:44:58,246][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:44:58,248][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-192
[2025-03-23 15:44:58,581][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 15:44:58,582][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 15:51:14,206][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:51:14,208][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:51:14,209][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:51:14,209][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:51:37,193][transformers][INFO] - {'accuracy': 0.6590909090909091, 'RMSE': 24.863262420322442, 'QWK': 0.6629280993390747, 'HDIV': 0.0, 'Macro_F1': 0.45253388586721915, 'Micro_F1': 0.6590909090909091, 'Weighted_F1': 0.6151236530024409, 'Macro_F1_(ignoring_nan)': np.float64(0.565667357334024)}
[2025-03-23 15:51:37,193][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:51:37,196][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-224
[2025-03-23 15:51:37,529][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 15:51:37,530][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 15:51:38,236][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-192] due to args.save_total_limit
[2025-03-23 15:57:53,116][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 15:57:53,118][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 15:57:53,118][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 15:57:53,118][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 15:58:16,064][transformers][INFO] - {'accuracy': 0.553030303030303, 'RMSE': 28.06917861068948, 'QWK': 0.6053357865685374, 'HDIV': 0.0, 'Macro_F1': 0.38989173236304836, 'Micro_F1': 0.553030303030303, 'Weighted_F1': 0.516636553528003, 'Macro_F1_(ignoring_nan)': np.float64(0.48736466545381046)}
[2025-03-23 15:58:16,065][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 15:58:16,068][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-256
[2025-03-23 15:58:26,827][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-224] due to args.save_total_limit
[2025-03-23 16:04:41,848][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 16:04:41,850][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 16:04:41,850][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 16:04:41,850][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 16:05:04,713][transformers][INFO] - {'accuracy': 0.6742424242424242, 'RMSE': 26.285149626910837, 'QWK': 0.6787909836065573, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.49290131869862563, 'Micro_F1': 0.6742424242424242, 'Weighted_F1': 0.6694937343413603, 'Macro_F1_(ignoring_nan)': np.float64(0.616126648373282)}
[2025-03-23 16:05:04,714][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 16:05:04,716][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-288
[2025-03-23 16:05:05,401][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 16:05:05,401][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 16:05:06,147][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-256] due to args.save_total_limit
[2025-03-23 16:11:21,262][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 16:11:21,264][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 16:11:21,264][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 16:11:21,264][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 16:11:44,181][transformers][INFO] - {'accuracy': 0.6742424242424242, 'RMSE': 26.285149626910837, 'QWK': 0.6179935012185216, 'HDIV': 0.007575757575757569, 'Macro_F1': 0.4704739704739705, 'Micro_F1': 0.6742424242424242, 'Weighted_F1': 0.6576546349273622, 'Macro_F1_(ignoring_nan)': np.float64(0.5880924630924631)}
[2025-03-23 16:11:44,181][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 16:11:44,184][transformers.trainer][INFO] - Saving model checkpoint to /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-320
[2025-03-23 16:11:44,464][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 16:11:44,465][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 16:11:45,140][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-288] due to args.save_total_limit
[2025-03-23 16:11:45,183][transformers.trainer][INFO] -
Training completed. Do not forget to share your model on huggingface.co/models =)
[2025-03-23 16:11:45,183][transformers.trainer][INFO] - Loading best model from /workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-160 (score: 0.6984548928393421).
[2025-03-23 16:11:45,419][transformers.trainer][INFO] - Deleting older checkpoint [/workspace/jbcs2025/outputs/2025-03-23/15-04-12/results/phi4-balanced/C1/checkpoint-320] due to args.save_total_limit
[2025-03-23 16:11:45,470][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 16:11:45,471][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 16:11:45,471][transformers.trainer][INFO] - Num examples = 132
[2025-03-23 16:11:45,471][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 16:12:08,400][transformers][INFO] - {'accuracy': 0.6742424242424242, 'RMSE': 25.81988897471611, 'QWK': 0.6984548928393421, 'HDIV': 0.0, 'Macro_F1': 0.5295884500497843, 'Micro_F1': 0.6742424242424242, 'Weighted_F1': 0.6732932544830984, 'Macro_F1_(ignoring_nan)': np.float64(0.6619855625622303)}
[2025-03-23 16:12:08,402][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 16:12:08,403][__main__][INFO] - Training completed successfully.
[2025-03-23 16:12:08,403][__main__][INFO] - Running on Test
[2025-03-23 16:12:08,403][transformers.trainer][INFO] - The following columns in the evaluation set don't have a corresponding argument in `PeftModelForSequenceClassification.forward` and have been ignored: id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades. If id, essay_text, essay_year, id_prompt, prompt, reference, supporting_text, grades are not expected by `PeftModelForSequenceClassification.forward`, you can safely ignore this message.
[2025-03-23 16:12:08,405][transformers.trainer][INFO] -
***** Running Evaluation *****
[2025-03-23 16:12:08,405][transformers.trainer][INFO] - Num examples = 138
[2025-03-23 16:12:08,405][transformers.trainer][INFO] - Batch size = 16
[2025-03-23 16:12:32,927][transformers][INFO] - {'accuracy': 0.6159420289855072, 'RMSE': 27.240208984279956, 'QWK': 0.6748637903107053, 'HDIV': 0.007246376811594235, 'Macro_F1': 0.46569434642829133, 'Micro_F1': 0.6159420289855072, 'Weighted_F1': 0.6184188545576659, 'Macro_F1_(ignoring_nan)': np.float64(0.5821179330353642)}
[2025-03-23 16:12:32,927][tensorboardX.summary][INFO] - Summary name eval/Macro_F1_(ignoring_nan) is illegal; using eval/Macro_F1__ignoring_nan_ instead.
[2025-03-23 16:12:32,930][transformers.trainer][INFO] - Saving model checkpoint to ./results/phi4-balanced/C1/best_model
[2025-03-23 16:12:33,185][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 16:12:33,185][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 16:12:33,735][__main__][INFO] - Fine Tuning Finished.