<!-- 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. -->
wav2vec2-base-timit-demo-colab6
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9394
- Wer: 0.5282
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 60
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.3117 | 7.35 | 500 | 3.1548 | 1.0 |
| 1.6732 | 14.71 | 1000 | 0.8857 | 0.6561 |
| 0.5267 | 22.06 | 1500 | 0.7931 | 0.6018 |
| 0.2951 | 29.41 | 2000 | 0.8152 | 0.5816 |
| 0.2013 | 36.76 | 2500 | 0.9060 | 0.5655 |
| 0.1487 | 44.12 | 3000 | 0.9201 | 0.5624 |
| 0.1189 | 51.47 | 3500 | 0.9394 | 0.5412 |
| 0.1004 | 58.82 | 4000 | 0.9394 | 0.5282 |
Framework versions
- Transformers 4.11.3
- Pytorch 1.11.0+cu113
- Datasets 1.18.3
- Tokenizers 0.10.3
