whisper-small-common-voice-3
This model is a fine-tuned version of lejonck/whisper-small-common-voice-2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 0.1207
- Wer: 0.2481
- Cer: 0.3645
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 |
|---|---|---|---|---|---|
| 0.2347 | 1.0 | 1000 | 0.1108 | 0.3383 | 0.3745 |
| 0.0761 | 2.0 | 2000 | 0.1207 | 0.2481 | 0.3645 |
| 0.0244 | 3.0 | 3000 | 0.1340 | 0.4093 | 0.3905 |
| 0.0076 | 4.0 | 4000 | 0.1434 | 0.4784 | 0.4075 |
| 0.0018 | 5.0 | 5000 | 0.1585 | 0.3921 | 0.3755 |
| 0.0035 | 6.0 | 6000 | 0.1639 | 0.4190 | 0.3841 |
| 0.0004 | 7.0 | 7000 | 0.1693 | 0.3445 | 0.3757 |
Framework versions
- Transformers 4.55.2
- Pytorch 2.7.0+cu126
- Datasets 2.19.1
- Tokenizers 0.21.4