whispertiny_buah25sil_v1_6
This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0698
- Wer: 5.5820
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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0383 | 0.2286 | 1000 | 0.0742 | 8.8113 |
| 0.0287 | 0.4571 | 2000 | 0.0605 | 6.2125 |
| 0.0219 | 0.6857 | 3000 | 0.0548 | 5.2745 |
| 0.0135 | 0.9143 | 4000 | 0.0696 | 5.6205 |
| 0.0069 | 1.1429 | 5000 | 0.0698 | 5.5820 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 4.0.0
- Tokenizers 0.19.1