<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
whisper-medium-ft-1000
This model is a fine-tuned version of openai/whisper-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2696
- Wer: 42.8039
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.5191 | 1.0 | 63 | 9.1980 | 100.0 |
| 4.0738 | 2.0 | 126 | 0.3468 | 18.3978 |
| 0.2178 | 3.0 | 189 | 0.2657 | 42.5710 |
| 0.0678 | 4.0 | 252 | 0.2659 | 18.0717 |
| 0.0247 | 5.0 | 315 | 0.2692 | 48.3931 |
| 0.0077 | 6.0 | 378 | 0.2696 | 42.8039 |
Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.13.3
