openai/whisper-medium
This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3896
- Wer: 200.1910
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: 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: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2328 | 0.12 | 500 | 0.2655 | 301.5949 |
| 0.1838 | 1.11 | 1000 | 0.2496 | 286.1977 |
| 0.1757 | 2.1 | 1500 | 0.2563 | 118.9213 |
| 0.0254 | 3.09 | 2000 | 0.2992 | 237.0841 |
| 0.0282 | 4.07 | 2500 | 0.3342 | 125.1999 |
| 0.0229 | 5.06 | 3000 | 0.3502 | 268.7414 |
| 0.0027 | 6.05 | 3500 | 0.3918 | 107.5536 |
| 0.003 | 7.03 | 4000 | 0.3896 | 200.1910 |
Framework versions
- Transformers 4.27.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.9.1.dev0
- Tokenizers 0.13.2