whisper-tiny-finetuned-minds14-enUS_2
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: 0.7508
- Wer Ortho: 0.3356
- Wer: 0.3394
- Cer: 0.2613
- Cer Ortho: 0.2623
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: 5e-05
- train_batch_size: 32
- 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: 50
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer | Cer | Cer Ortho |
|---|---|---|---|---|---|---|---|
| 0.0136 | 7.14 | 100 | 0.6142 | 0.3362 | 0.3388 | 0.2587 | 0.2614 |
| 0.0009 | 14.29 | 200 | 0.6704 | 0.3288 | 0.3300 | 0.2515 | 0.2534 |
| 0.0011 | 21.43 | 300 | 0.6858 | 0.3054 | 0.3093 | 0.2363 | 0.2374 |
| 0.0005 | 28.57 | 400 | 0.7081 | 0.3455 | 0.3477 | 0.2699 | 0.2711 |
| 0.0004 | 35.71 | 500 | 0.7191 | 0.3467 | 0.3501 | 0.2727 | 0.2736 |
| 0.0001 | 42.86 | 600 | 0.7337 | 0.3405 | 0.3447 | 0.2652 | 0.2662 |
| 0.0001 | 50.0 | 700 | 0.7418 | 0.3393 | 0.3430 | 0.2636 | 0.2645 |
| 0.0001 | 57.14 | 800 | 0.7466 | 0.3387 | 0.3424 | 0.2634 | 0.2644 |
| 0.0001 | 64.29 | 900 | 0.7496 | 0.3350 | 0.3388 | 0.2604 | 0.2614 |
| 0.0001 | 71.43 | 1000 | 0.7508 | 0.3356 | 0.3394 | 0.2613 | 0.2623 |
Framework versions
- Transformers 4.32.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3