wav2vec2_test
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: 91.1661
- Wer: 0.5714
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: 100
- num_epochs: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 11.9459 | 100.0 | 100 | 46.9901 | 1.0 |
| 3.2175 | 200.0 | 200 | 73.0950 | 1.0 |
| 1.8117 | 300.0 | 300 | 78.4884 | 0.6735 |
| 1.3694 | 400.0 | 400 | 84.0168 | 0.6327 |
| 1.1392 | 500.0 | 500 | 85.2083 | 0.5918 |
| 0.979 | 600.0 | 600 | 88.9109 | 0.5918 |
| 0.8917 | 700.0 | 700 | 89.0310 | 0.5918 |
| 0.8265 | 800.0 | 800 | 90.0659 | 0.6122 |
| 0.769 | 900.0 | 900 | 91.8476 | 0.5714 |
| 0.7389 | 1000.0 | 1000 | 91.1661 | 0.5714 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.10.2
- Datasets 1.18.3
- Tokenizers 0.11.6