CS221-deberta-v3-large-finetuned-semeval
This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3471
- F1: 0.7809
- Roc Auc: 0.8338
- Accuracy: 0.4856
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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Roc Auc | Accuracy |
|---|---|---|---|---|---|---|
| 0.5136 | 1.0 | 139 | 0.4650 | 0.6413 | 0.7342 | 0.2996 |
| 0.3853 | 2.0 | 278 | 0.3781 | 0.7090 | 0.7789 | 0.4097 |
| 0.2579 | 3.0 | 417 | 0.3471 | 0.7809 | 0.8338 | 0.4856 |
| 0.1682 | 4.0 | 556 | 0.4160 | 0.7546 | 0.8126 | 0.4657 |
| 0.0857 | 5.0 | 695 | 0.4492 | 0.7754 | 0.8300 | 0.4946 |
| 0.0485 | 6.0 | 834 | 0.5044 | 0.7741 | 0.8313 | 0.4675 |
Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0