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: 0.3246
- Wer: 341.9230
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 |
|---|---|---|---|---|
| 0.2032 | 0.12 | 500 | 0.2243 | 493.2750 |
| 0.1192 | 1.1 | 1000 | 0.2127 | 424.6297 |
| 0.1109 | 2.08 | 1500 | 0.2237 | 351.5590 |
| 0.042 | 3.06 | 2000 | 0.2460 | 165.9201 |
| 0.0262 | 4.04 | 2500 | 0.2909 | 231.2864 |
| 0.0139 | 5.02 | 3000 | 0.3042 | 350.0223 |
| 0.0084 | 6.0 | 3500 | 0.3247 | 327.0151 |
| 0.0023 | 6.13 | 4000 | 0.3246 | 341.9230 |
Framework versions
- Transformers 4.27.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.9.1.dev0
- Tokenizers 0.13.2