whisper-base.en-atcosim
This model is a fine-tuned version of openai/whisper-base.en on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0634
- Wer: 4.1443
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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- num_epochs: 100
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3402 | 8.33 | 500 | 0.0540 | 5.1398 |
| 0.0011 | 16.67 | 1000 | 0.0557 | 4.2693 |
| 0.0003 | 25.0 | 1500 | 0.0571 | 3.9128 |
| 0.0002 | 33.33 | 2000 | 0.0583 | 3.7553 |
| 0.0001 | 41.67 | 2500 | 0.0594 | 3.5840 |
| 0.0001 | 50.0 | 3000 | 0.0603 | 3.4729 |
| 0.0001 | 58.33 | 3500 | 0.0613 | 3.3571 |
| 0.0001 | 66.67 | 4000 | 0.0619 | 3.3108 |
| 0.0 | 75.0 | 4500 | 0.0625 | 4.2369 |
| 0.0 | 83.33 | 5000 | 0.0630 | 4.2184 |
| 0.0 | 91.67 | 5500 | 0.0633 | 4.1489 |
| 0.0 | 100.0 | 6000 | 0.0634 | 4.1443 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2
- Datasets 2.15.0
- Tokenizers 0.15.0