whisper-base-common_voice_17_0-id
This model is a fine-tuned version of openai/whisper-base on the mozilla-foundation/common_voice_17_0 id dataset. It achieves the following results on the evaluation set:
- Loss: 0.1441
- Wer: 0.1184
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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 20000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3523 | 0.4229 | 1000 | 0.3129 | 0.2365 |
| 0.3002 | 0.8458 | 2000 | 0.2391 | 0.1964 |
| 0.1718 | 1.2688 | 3000 | 0.2049 | 0.1659 |
| 0.1537 | 1.6917 | 4000 | 0.1817 | 0.1516 |
| 0.0807 | 2.1146 | 5000 | 0.1643 | 0.1499 |
| 0.089 | 2.5375 | 6000 | 0.1562 | 0.1348 |
| 0.0883 | 2.9605 | 7000 | 0.1452 | 0.1268 |
| 0.0368 | 3.3834 | 8000 | 0.1446 | 0.1324 |
| 0.0463 | 3.8063 | 9000 | 0.1401 | 0.1286 |
| 0.0278 | 4.2292 | 10000 | 0.1436 | 0.1181 |
| 0.0157 | 4.6521 | 11000 | 0.1406 | 0.1125 |
| 0.0201 | 5.0751 | 12000 | 0.1392 | 0.1144 |
| 0.0121 | 5.4980 | 13000 | 0.1405 | 0.1129 |
| 0.0074 | 5.9209 | 14000 | 0.1385 | 0.1195 |
| 0.0064 | 6.3438 | 15000 | 0.1410 | 0.1115 |
| 0.0066 | 6.7668 | 16000 | 0.1415 | 0.1184 |
| 0.0029 | 7.1897 | 17000 | 0.1426 | 0.1190 |
| 0.0024 | 7.6126 | 18000 | 0.1429 | 0.1178 |
| 0.0021 | 8.0355 | 19000 | 0.1434 | 0.1180 |
| 0.0018 | 8.4584 | 20000 | 0.1441 | 0.1184 |
Framework versions
- Transformers 4.42.0.dev0
- Pytorch 2.1.0
- Datasets 2.19.1
- Tokenizers 0.19.1