whisper-small-E10_freq_pause_speed
This model is a fine-tuned version of openai/whisper-small on the aihub old adult freq speed pause changed dataset. It achieves the following results on the evaluation set:
- Loss: 0.2281
- Cer: 6.1325
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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: 50
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.269 | 0.1289 | 100 | 0.2856 | 7.1898 |
| 0.1585 | 0.2579 | 200 | 0.2550 | 6.8080 |
| 0.1272 | 0.3868 | 300 | 0.2569 | 7.0254 |
| 0.1285 | 0.5158 | 400 | 0.2428 | 6.9196 |
| 0.11 | 0.6447 | 500 | 0.2463 | 6.7963 |
| 0.0981 | 0.7737 | 600 | 0.2459 | 6.7375 |
| 0.0998 | 0.9026 | 700 | 0.2378 | 6.3792 |
| 0.0378 | 1.0309 | 800 | 0.2264 | 6.0150 |
| 0.0281 | 1.1599 | 900 | 0.2285 | 5.9152 |
| 0.0344 | 1.2888 | 1000 | 0.2311 | 6.1560 |
| 0.0336 | 1.4178 | 1100 | 0.2295 | 6.1384 |
| 0.0364 | 1.5467 | 1200 | 0.2306 | 6.1325 |
| 0.0347 | 1.6757 | 1300 | 0.2300 | 6.1266 |
| 0.0317 | 1.8046 | 1400 | 0.2282 | 6.1208 |
| 0.036 | 1.9336 | 1500 | 0.2281 | 6.1325 |
Framework versions
- Transformers 4.47.0.dev0
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3