whisper-tiny-minds14-en-us
This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5930
- Wer: 0.3165
- Wer Ortho: 0.3448
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: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Wer Ortho |
|---|---|---|---|---|---|
| 1.7385 | 1.0 | 29 | 0.7435 | 0.3751 | 0.4170 |
| 0.5314 | 2.0 | 58 | 0.5790 | 0.3442 | 0.3757 |
| 0.3243 | 3.0 | 87 | 0.5545 | 0.3393 | 0.3732 |
| 0.2605 | 4.0 | 116 | 0.5519 | 0.3374 | 0.3701 |
| 0.2126 | 5.0 | 145 | 0.5595 | 0.3251 | 0.3578 |
| 0.1212 | 6.0 | 174 | 0.5680 | 0.3251 | 0.3547 |
| 0.1061 | 7.0 | 203 | 0.5810 | 0.3239 | 0.3523 |
| 0.0719 | 8.0 | 232 | 0.5878 | 0.3196 | 0.3486 |
| 0.0834 | 9.0 | 261 | 0.5930 | 0.3165 | 0.3448 |
| 0.0464 | 10.0 | 290 | 0.5955 | 0.3165 | 0.3442 |
Framework versions
- Transformers 4.57.6
- Pytorch 2.11.0+cu128
- Datasets 3.6.0
- Tokenizers 0.22.2