This model was trained from scratch on the librispeech_asr dataset. It achieves the following results on the evaluation set:
- Loss: 6.5541
- Wer: 1.9877
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: 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
- num_epochs: 3.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.9364 | 0.28 | 500 | 6.3613 | 1.9833 |
| 1.941 | 0.56 | 1000 | 5.6974 | 1.9746 |
| 2.3312 | 0.84 | 1500 | 5.6979 | 1.7345 |
| 2.8004 | 1.12 | 2000 | 6.0436 | 1.6787 |
| 3.0003 | 1.4 | 2500 | 6.0955 | 1.7625 |
| 2.9677 | 1.68 | 3000 | 6.2841 | 1.6731 |
| 2.2759 | 1.96 | 3500 | 6.3094 | 1.7494 |
| 2.2989 | 2.24 | 4000 | 6.9891 | 1.9115 |
| 1.8814 | 2.52 | 4500 | 6.9818 | 1.9832 |
| 2.658 | 2.8 | 5000 | 6.5541 | 1.9877 |
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0