| 2026-04-28 08:56:48,286 - INFO - ===== Training Configuration ===== |
| 2026-04-28 08:56:48,288 - INFO - model_name : microsoft/graphcodebert-base |
| 2026-04-28 08:56:48,289 - INFO - output_dir : output_checkpoints/graphcodebert-best/ |
| 2026-04-28 08:56:48,290 - INFO - num_epochs : 2 |
| 2026-04-28 08:56:48,290 - INFO - max_steps : -1 |
| 2026-04-28 08:56:48,291 - INFO - batch_size : 512 |
| 2026-04-28 08:56:48,292 - INFO - learning_rate : 1e-06 |
| 2026-04-28 08:56:48,293 - INFO - max_length : 512 |
| 2026-04-28 08:56:48,294 - INFO - num_labels : 2 |
| 2026-04-28 08:56:48,295 - INFO - use_wandb : True |
| 2026-04-28 08:56:48,295 - INFO - freeze_base : True |
| 2026-04-28 08:56:48,296 - INFO - loss_type : r-drop |
| 2026-04-28 08:56:48,297 - INFO - focal_alpha : 1.0 |
| 2026-04-28 08:56:48,298 - INFO - focal_gamma : 2.0 |
| 2026-04-28 08:56:48,299 - INFO - r_drop_alpha : 6.0 |
| 2026-04-28 08:56:48,300 - INFO - infonce_temperature : 0.07 |
| 2026-04-28 08:56:48,301 - INFO - infonce_weight : 0.5 |
| 2026-04-28 08:56:48,301 - INFO - seed : 42 |
| 2026-04-28 08:56:48,303 - INFO - wandb_run_name : graphcodebert-rdrop-best |
| 2026-04-28 08:56:48,303 - INFO - resume_from_checkpoint : checkpoints/graphcodebert-base-lowLR-highBatchSize/checkpoint-1022 |
| 2026-04-28 08:56:48,304 - INFO - save_steps : 100 |
| 2026-04-28 08:56:48,305 - INFO - eval_steps : 20 |
| 2026-04-28 08:56:48,306 - INFO - logging_steps : 2 |
| 2026-04-28 08:56:48,307 - INFO - label_smoothing : 0.5 |
| 2026-04-28 08:56:48,307 - INFO - adversarial_epsilon : 0.5 |
| 2026-04-28 08:56:48,308 - INFO - use_swa : False |
| 2026-04-28 08:56:48,309 - INFO - swa_start_epoch : 0 |
| 2026-04-28 08:56:48,309 - INFO - swa_lr : 1e-05 |
| 2026-04-28 08:56:48,310 - INFO - data_augmentation : True |
| 2026-04-28 08:56:48,311 - INFO - aug_rename_prob : 0.8 |
| 2026-04-28 08:56:48,313 - INFO - aug_format_prob : 0.8 |
| 2026-04-28 08:56:48,313 - INFO - mixup_alpha : 1.0 |
| 2026-04-28 08:56:48,314 - INFO - low_pass_keep_ratio : 0.5 |
| 2026-04-28 08:56:48,315 - INFO - freq_consistency_weight : 0.5 |
| 2026-04-28 08:56:48,315 - INFO - hidden_dropout_prob : 0.3 |
| 2026-04-28 08:56:48,316 - INFO - attention_probs_dropout_prob : 0.3 |
| 2026-04-28 08:56:48,317 - INFO - classifier_dropout : 0.3 |
| 2026-04-28 08:56:48,318 - INFO - ================================= |
| 2026-04-28 08:56:49,588 - INFO - Model placed on cuda |
| 2026-04-28 08:56:49,593 - INFO - ===== Model Architecture ===== |
| 2026-04-28 08:56:49,595 - 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=0.3, inplace=False) |
| (out_proj): Linear(in_features=768, out_features=2, bias=True) |
| ) |
| ) |
| 2026-04-28 08:56:49,597 - INFO - ===== Parameter Summary ===== |
| 2026-04-28 08:56:49,598 - INFO - Total Parameters: 124,647,170 |
| 2026-04-28 08:56:49,599 - INFO - Trainable Parameters: 592,130 |
| 2026-04-28 08:56:49,599 - INFO - Non-trainable Parameters: 124,055,040 |
| 2026-04-28 08:56:49,600 - INFO - ===== Tokenizer Summary ===== |
| 2026-04-28 08:56:49,615 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>'] |
| 2026-04-28 08:56:49,617 - INFO - ===== End of Architecture Log ===== |
| 2026-04-28 08:56:49,617 - INFO - Data augmentation enabled (rename=0.8, format=0.8) |
| 2026-04-28 08:57:21,966 - INFO - === Starting training with MixCode + FFT low-pass consistency === |
| 2026-04-28 10:55:30,390 - INFO - Training interrupted by user. |
|
|
