<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
This model was trained from scratch on the librispeech_asr dataset. It achieves the following results on the evaluation set:
- Loss: 0.8384
- Wer: 0.1367
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 20.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.2245 | 1.68 | 1500 | 6.1442 | 1.5986 |
| 5.4521 | 3.36 | 3000 | 5.4335 | 1.6439 |
| 3.3659 | 5.04 | 4500 | 3.6455 | 0.6503 |
| 1.5724 | 6.73 | 6000 | 2.3554 | 0.3386 |
| 1.4759 | 8.41 | 7500 | 1.7423 | 0.2889 |
| 1.0826 | 10.09 | 9000 | 1.3818 | 0.2209 |
| 0.6769 | 11.77 | 10500 | 1.1268 | 0.1737 |
| 0.7348 | 13.45 | 12000 | 0.9990 | 0.1575 |
| 0.5419 | 15.13 | 13500 | 0.9435 | 0.1560 |
| 0.4212 | 16.82 | 15000 | 0.8678 | 0.1405 |
| 0.3805 | 18.5 | 16500 | 0.8384 | 0.1367 |
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
