HamdanXI/wav2vec2-base-myst-new
automatic speech recognitiontransformerstransformerssafetensorswav2vec2automatic-speech-recognitiongenerated_from_trainerbase_model:facebook/wav2vec2-baseapache-2.0
HuggingFace408.6K
wav2vec2-base-myst-new
This model is a fine-tuned version of facebook/wav2vec2-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4262
- Wer: 0.1249
- Cer: 0.0583
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.06
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.5255 | 1.1615 | 2000 | 0.4068 | 0.2182 | 0.0884 |
| 0.358 | 2.3229 | 4000 | 0.3476 | 0.1657 | 0.0713 |
| 0.3055 | 3.4844 | 6000 | 0.3039 | 0.1564 | 0.0663 |
| 0.2571 | 4.6459 | 8000 | 0.2945 | 0.1472 | 0.0641 |
| 0.2373 | 5.8073 | 10000 | 0.3080 | 0.1457 | 0.0635 |
| 0.2277 | 6.9688 | 12000 | 0.3035 | 0.1370 | 0.0619 |
| 0.184 | 8.1301 | 14000 | 0.3264 | 0.1336 | 0.0603 |
| 0.155 | 9.2916 | 16000 | 0.3322 | 0.1348 | 0.0614 |
| 0.1459 | 10.4530 | 18000 | 0.3464 | 0.1340 | 0.0617 |
| 0.1396 | 11.6145 | 20000 | 0.3306 | 0.1330 | 0.0610 |
| 0.1288 | 12.7760 | 22000 | 0.3563 | 0.1294 | 0.0595 |
| 0.1123 | 13.9374 | 24000 | 0.3605 | 0.1294 | 0.0598 |
| 0.1061 | 15.0987 | 26000 | 0.3896 | 0.1287 | 0.0595 |
| 0.0938 | 16.2602 | 28000 | 0.3904 | 0.1274 | 0.0591 |
| 0.0851 | 17.4217 | 30000 | 0.4189 | 0.1252 | 0.0585 |
| 0.0858 | 18.5831 | 32000 | 0.4145 | 0.1260 | 0.0587 |
| 0.076 | 19.7446 | 34000 | 0.4262 | 0.1249 | 0.0583 |
Framework versions
- Transformers 4.57.0
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.22.1