whisper-small-ptbr-cv-final3
This model is a fine-tuned version of lejonck/whisper-small-ptbr-cv-final2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3336
- Wer: 0.2704
- Cer: 0.3462
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: 8
- eval_batch_size: 2
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH 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: 12
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.3003 | 1.0 | 500 | 0.3107 | 0.2723 | 0.3453 |
| 0.1105 | 2.0 | 1000 | 0.3293 | 0.2708 | 0.3486 |
| 0.0365 | 3.0 | 1500 | 0.3336 | 0.2757 | 0.3492 |
| 0.0127 | 4.0 | 2000 | 0.4017 | 0.2723 | 0.3485 |
| 0.0041 | 5.0 | 2500 | 0.4252 | 0.2723 | 0.3460 |
| 0.0053 | 6.0 | 3000 | 0.4572 | 0.2775 | 0.3529 |
| 0.0007 | 7.0 | 3500 | 0.4898 | 0.2742 | 0.3500 |
| 0.0006 | 8.0 | 4000 | 0.4761 | 0.2719 | 0.3484 |
Framework versions
- Transformers 4.55.1
- Pytorch 2.7.0+cu126
- Datasets 2.19.1
- Tokenizers 0.21.4