codebert-code-clone-detector
This model is a fine-tuned version of microsoft/codebert-base on a Code Clone Benchmark dataset. See this github repository for more information. It achieves the following results on the evaluation set:
- Loss: 0.3452
- Accuracy: 0.9525
- Precision: 0.9544
- Recall: 0.9496
- F1: 0.9520
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.3416 | 0.49 | 33 | 0.1724 | 0.9417 | 0.9828 | 0.9048 | 0.9421 |
| 0.221 | 0.97 | 66 | 0.2768 | 0.925 | 1.0 | 0.8571 | 0.9231 |
| 0.0929 | 1.46 | 99 | 0.2469 | 0.9583 | 1.0 | 0.9206 | 0.9587 |
| 0.1696 | 1.94 | 132 | 0.2142 | 0.95 | 0.9524 | 0.9524 | 0.9524 |
| 0.0818 | 2.43 | 165 | 0.4142 | 0.925 | 1.0 | 0.8571 | 0.9231 |
| 0.0676 | 2.91 | 198 | 0.3539 | 0.9333 | 0.9508 | 0.9206 | 0.9355 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2