whisper-small-clean-1e5-10epochs
This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3349
- Wer: 58.3694
- Cer: 20.4568
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 2.2896 | 1.2675 | 100 | 0.5705 | 80.6671 | 33.0952 |
| 0.8333 | 2.5350 | 200 | 0.2642 | 68.6844 | 23.7372 |
| 0.5514 | 3.8025 | 300 | 0.2351 | 61.2724 | 20.9197 |
| 0.2468 | 5.0637 | 400 | 0.2479 | 60.7783 | 21.0908 |
| 0.1454 | 6.3312 | 500 | 0.2713 | 59.0488 | 20.0443 |
| 0.0654 | 7.5987 | 600 | 0.3037 | 59.5429 | 20.3461 |
| 0.0353 | 8.8662 | 700 | 0.3349 | 58.3694 | 20.4568 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 4.8.3
- Tokenizers 0.22.2