dzungpham/graphcodebert-code-classification

12026-04-24 17:10:51,664 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-base-lowLR-highBatchSize/checkpoint-40022026-04-24 17:10:51,869 - INFO - ===== Model Architecture =====32026-04-24 17:10:51,872 - INFO - 4RobertaForSequenceClassification(5  (roberta): RobertaModel(6    (embeddings): RobertaEmbeddings(7      (word_embeddings): Embedding(50265, 768, padding_idx=1)8      (position_embeddings): Embedding(514, 768, padding_idx=1)9      (token_type_embeddings): Embedding(1, 768)10      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)11      (dropout): Dropout(p=0.3, inplace=False)12    )13    (encoder): RobertaEncoder(14      (layer): ModuleList(15        (0-11): 12 x RobertaLayer(16          (attention): RobertaAttention(17            (self): RobertaSdpaSelfAttention(18              (query): Linear(in_features=768, out_features=768, bias=True)19              (key): Linear(in_features=768, out_features=768, bias=True)20              (value): Linear(in_features=768, out_features=768, bias=True)21              (dropout): Dropout(p=0.3, inplace=False)22            )23            (output): RobertaSelfOutput(24              (dense): Linear(in_features=768, out_features=768, bias=True)25              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)26              (dropout): Dropout(p=0.3, inplace=False)27            )28          )29          (intermediate): RobertaIntermediate(30            (dense): Linear(in_features=768, out_features=3072, bias=True)31            (intermediate_act_fn): GELUActivation()32          )33          (output): RobertaOutput(34            (dense): Linear(in_features=3072, out_features=768, bias=True)35            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)36            (dropout): Dropout(p=0.3, inplace=False)37          )38        )39      )40    )41  )42  (classifier): RobertaClassificationHead(43    (dense): Linear(in_features=768, out_features=768, bias=True)44    (dropout): Dropout(p=0.3, inplace=False)45    (out_proj): Linear(in_features=768, out_features=2, bias=True)46  )47)482026-04-24 17:10:51,874 - INFO - ===== Parameter Summary =====492026-04-24 17:10:51,876 - INFO - Total Parameters:         124,647,170502026-04-24 17:10:51,877 - INFO - Trainable Parameters:     124,647,170512026-04-24 17:10:51,879 - INFO - Non-trainable Parameters: 0522026-04-24 17:10:51,881 - INFO - ===== Tokenizer Summary =====532026-04-24 17:10:51,896 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']542026-04-24 17:10:51,898 - INFO - ===== End of Architecture Log =====552026-04-24 17:10:52,067 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)562026-04-24 17:10:52,562 - INFO - Tokenizing dataset...572026-04-24 17:10:53,909 - INFO - Running inference on 1000 examples...582026-04-24 17:11:23,974 - INFO - Calculating classification metrics...592026-04-24 17:11:23,994 - INFO - ------------------------------602026-04-24 17:11:23,996 - INFO - METRICS FOR SPLIT: test612026-04-24 17:11:23,997 - INFO - Accuracy:  0.7400622026-04-24 17:11:23,999 - INFO - Precision: 0.6710632026-04-24 17:11:24,000 - INFO - Recall:    0.7400642026-04-24 17:11:24,001 - INFO - F1-Score:  0.6924652026-04-24 17:11:24,002 - INFO - ------------------------------662026-04-24 17:11:24,004 - INFO - Confusion Matrix:67[[716  61]68 [199  24]]692026-04-24 17:11:24,006 - INFO - ✅ Metrics saved to test/inference/graphcodebert-base-lowLR-highBatchSize/checkpoint-400/metrics.json702026-04-24 17:11:24,008 - INFO - ✅ Predictions saved to test/inference/graphcodebert-base-lowLR-highBatchSize/checkpoint-400/submission.csv71