wav2vec2-large-xlrs
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2309
- Wer: 0.2314
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: 32
- 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: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 4.6199 | 3.45 | 500 | 3.1926 | 1.0 |
| 2.5032 | 6.9 | 1000 | 0.6387 | 0.5685 |
| 0.4994 | 10.34 | 1500 | 0.2603 | 0.2971 |
| 0.2602 | 13.79 | 2000 | 0.2264 | 0.2562 |
| 0.1763 | 17.24 | 2500 | 0.2434 | 0.2450 |
| 0.1459 | 20.69 | 3000 | 0.2197 | 0.2370 |
| 0.1326 | 24.14 | 3500 | 0.2301 | 0.2333 |
| 0.1088 | 27.59 | 4000 | 0.2309 | 0.2314 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.1+cu116
- Datasets 1.18.3
- Tokenizers 0.13.2