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: 0.5021
- Wer: 11.9280
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 |
|---|---|---|---|---|
| 0.3007 | 1.01 | 500 | 0.3027 | 12.4842 |
| 0.0962 | 2.02 | 1000 | 0.3147 | 11.9450 |
| 0.0409 | 3.03 | 1500 | 0.3528 | 12.9437 |
| 0.0298 | 4.04 | 2000 | 0.4015 | 11.7048 |
| 0.0114 | 5.06 | 2500 | 0.4310 | 12.3353 |
| 0.0034 | 6.07 | 3000 | 0.4666 | 11.6630 |
| 0.0014 | 7.08 | 3500 | 0.5117 | 11.7505 |
| 0.0016 | 8.09 | 4000 | 0.5021 | 11.9280 |
Framework versions
- Transformers 4.27.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.9.1.dev0
- Tokenizers 0.13.2