whispertiny_kokoro_styletts_25sil_v1_0
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: 0.1003
- Wer: 8.1340
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: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0833 | 0.2286 | 1000 | 0.1509 | 17.2142 |
| 0.0734 | 0.4571 | 2000 | 0.1071 | 12.3193 |
| 0.0446 | 0.6857 | 3000 | 0.0965 | 10.9067 |
| 0.0396 | 0.9143 | 4000 | 0.1064 | 11.1367 |
| 0.0229 | 1.1429 | 5000 | 0.1071 | 10.2102 |
| 0.0166 | 1.3714 | 6000 | 0.0991 | 8.7516 |
| 0.0162 | 1.6 | 7000 | 0.0965 | 8.7319 |
| 0.0174 | 1.8286 | 8000 | 0.0942 | 8.3903 |
| 0.0131 | 2.0571 | 9000 | 0.0983 | 8.2194 |
| 0.0157 | 2.2857 | 10000 | 0.1003 | 8.1340 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 4.0.0
- Tokenizers 0.19.1