openai/whisper-large-v3-turbo
This model is a fine-tuned version of openai/whisper-large-v3-turbo on the common_voice_22_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4219
- Wer: 8.2359
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: 3.75e-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 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: 500
- training_steps: 100000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0243 | 10.8234 | 5000 | 0.2572 | 11.3209 |
| 0.012 | 21.6457 | 10000 | 0.2926 | 11.1214 |
| 0.007 | 32.4680 | 15000 | 0.3036 | 11.0386 |
| 0.0042 | 43.2904 | 20000 | 0.3139 | 10.8341 |
| 0.004 | 54.1127 | 25000 | 0.3344 | 11.1519 |
| 0.0043 | 64.9361 | 30000 | 0.3439 | 11.2584 |
| 0.0031 | 75.7584 | 35000 | 0.3333 | 10.4943 |
| 0.0023 | 86.5807 | 40000 | 0.3454 | 10.5746 |
| 0.0018 | 97.4030 | 45000 | 0.3443 | 9.9508 |
| 0.0012 | 108.2254 | 50000 | 0.3494 | 10.5188 |
| 0.0008 | 119.0477 | 55000 | 0.3521 | 10.1148 |
| 0.0005 | 129.8711 | 60000 | 0.3651 | 9.8663 |
| 0.0013 | 140.6934 | 65000 | 0.3651 | 10.3743 |
| 0.0002 | 151.5157 | 70000 | 0.3531 | 9.6043 |
| 0.0 | 162.3380 | 75000 | 0.3638 | 9.3287 |
| 0.0 | 173.1603 | 80000 | 0.3724 | 9.0532 |
| 0.0 | 183.9837 | 85000 | 0.3875 | 8.6906 |
| 0.0 | 194.8061 | 90000 | 0.4055 | 8.4083 |
| 0.0 | 205.6284 | 95000 | 0.4180 | 8.2730 |
| 0.0 | 216.4507 | 100000 | 0.4219 | 8.2359 |
Framework versions
- Transformers 4.52.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1