ID Optimized Whisper Small for Inspirasi
This model is a fine-tuned version of openai/whisper-small on the Extracted Youtube with self developed dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.3443
- Wer: 12.7922
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: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2816 | 0.9346 | 1000 | 0.2899 | 16.8831 |
| 0.1195 | 1.8692 | 2000 | 0.2892 | 15.5519 |
| 0.0453 | 2.8037 | 3000 | 0.2804 | 14.4481 |
| 0.0205 | 3.7383 | 4000 | 0.3164 | 13.9286 |
| 0.007 | 4.6729 | 5000 | 0.3248 | 13.5065 |
| 0.0045 | 5.6075 | 6000 | 0.3295 | 13.2468 |
| 0.0015 | 6.5421 | 7000 | 0.3354 | 12.6299 |
| 0.001 | 7.4766 | 8000 | 0.3398 | 12.5325 |
| 0.0008 | 8.4112 | 9000 | 0.3416 | 12.8247 |
| 0.0007 | 9.3458 | 10000 | 0.3443 | 12.7922 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.2.0a0+81ea7a4
- Datasets 3.6.0
- Tokenizers 0.21.1