2026-04-24 14:59:26,182 - INFO - ===== Training Configuration =====
2026-04-24 14:59:26,184 - INFO - model_name           : microsoft/graphcodebert-base
2026-04-24 14:59:26,185 - INFO - output_dir           : output_checkpoints/graphcodebert-ready-lmc/run_1
2026-04-24 14:59:26,186 - INFO - num_epochs           : 1000
2026-04-24 14:59:26,187 - INFO - max_steps            : -1
2026-04-24 14:59:26,189 - INFO - batch_size           : 16
2026-04-24 14:59:26,191 - INFO - learning_rate        : 5e-06
2026-04-24 14:59:26,192 - INFO - max_length           : 512
2026-04-24 14:59:26,193 - INFO - num_labels           : 2
2026-04-24 14:59:26,194 - INFO - use_wandb            : False
2026-04-24 14:59:26,196 - INFO - freeze_base          : True
2026-04-24 14:59:26,197 - INFO - loss_type            : r-drop
2026-04-24 14:59:26,199 - INFO - focal_alpha          : 1.0
2026-04-24 14:59:26,200 - INFO - focal_gamma          : 2.0
2026-04-24 14:59:26,201 - INFO - r_drop_alpha         : 4.0
2026-04-24 14:59:26,202 - INFO - infonce_temperature  : 0.07
2026-04-24 14:59:26,204 - INFO - infonce_weight       : 0.5
2026-04-24 14:59:26,206 - INFO - seed                 : 42
2026-04-24 14:59:26,208 - INFO - resume_from_checkpoint : None
2026-04-24 14:59:26,209 - INFO - label_smoothing      : 0.2
2026-04-24 14:59:26,211 - INFO - adversarial_epsilon  : 0.5
2026-04-24 14:59:26,212 - INFO - use_swa              : False
2026-04-24 14:59:26,213 - INFO - swa_start_epoch      : 0
2026-04-24 14:59:26,215 - INFO - swa_lr               : 1e-05
2026-04-24 14:59:26,216 - INFO - data_augmentation    : True
2026-04-24 14:59:26,218 - INFO - aug_rename_prob      : 0.6
2026-04-24 14:59:26,219 - INFO - aug_format_prob      : 0.6
2026-04-24 14:59:26,221 - INFO - mixup_alpha          : 0.8
2026-04-24 14:59:26,222 - INFO - low_pass_keep_ratio  : 0.5
2026-04-24 14:59:26,224 - INFO - freq_consistency_weight : 0.2
2026-04-24 14:59:26,225 - INFO - hidden_dropout_prob  : 0.3
2026-04-24 14:59:26,226 - INFO - attention_probs_dropout_prob : 0.3
2026-04-24 14:59:26,228 - INFO - classifier_dropout   : 0.2
2026-04-24 14:59:26,229 - INFO - =================================
2026-04-24 14:59:27,405 - INFO - Model placed on cuda
2026-04-24 14:59:27,409 - INFO - ===== Model Architecture =====
2026-04-24 14:59:27,412 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.2, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-24 14:59:27,414 - INFO - ===== Parameter Summary =====
2026-04-24 14:59:27,416 - INFO - Total Parameters:         124,647,170
2026-04-24 14:59:27,417 - INFO - Trainable Parameters:     592,130
2026-04-24 14:59:27,419 - INFO - Non-trainable Parameters: 124,055,040
2026-04-24 14:59:27,421 - INFO - ===== Tokenizer Summary =====
2026-04-24 14:59:27,436 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-24 14:59:27,438 - INFO - ===== End of Architecture Log =====
2026-04-24 14:59:27,441 - INFO - Data augmentation enabled (rename=0.6, format=0.6)