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: 0.3853
- Wer: 10.4258
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 |
|---|---|---|---|---|
| 0.2536 | 0.12 | 500 | 0.2608 | 11.8586 |
| 0.3687 | 1.1 | 1000 | 0.2578 | 11.4576 |
| 0.1522 | 2.07 | 1500 | 0.2613 | 12.7949 |
| 0.0387 | 3.05 | 2000 | 0.2952 | 10.9378 |
| 0.014 | 4.02 | 2500 | 0.3271 | 10.6813 |
| 0.0186 | 4.14 | 3000 | 0.3389 | 10.3970 |
| 0.0057 | 5.12 | 3500 | 0.3670 | 10.6380 |
| 0.0108 | 6.09 | 4000 | 0.3853 | 10.4258 |
Framework versions
- Transformers 4.29.0
- Pytorch 1.14.0a0+44dac51
- Datasets 2.12.0
- Tokenizers 0.13.3