Whisper Medium Portuguese
This model is a fine-tuned version of openai/whisper-medium on the mozilla-foundation/common_voice_13_0 pt dataset. It achieves the following results on the evaluation set:
- Loss: 0.1753
- Wer: 6.3319
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-06
- train_batch_size: 64
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0999 | 3.52 | 1000 | 0.1753 | 6.3319 |
| 0.0436 | 7.04 | 2000 | 0.2027 | 6.5521 |
| 0.0113 | 10.56 | 3000 | 0.3135 | 6.7361 |
| 0.0041 | 14.08 | 4000 | 0.3616 | 6.8889 |
| 0.0026 | 17.61 | 5000 | 0.3908 | 7.0565 |
| 0.0016 | 21.13 | 6000 | 0.4078 | 7.1419 |
| 0.0013 | 24.65 | 7000 | 0.4227 | 7.1534 |
| 0.001 | 28.17 | 8000 | 0.4343 | 7.1764 |
| 0.0008 | 31.69 | 9000 | 0.4424 | 7.2076 |
| 0.0008 | 35.21 | 10000 | 0.4464 | 7.2224 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.15.1