wav2vec2-large-xls-r-300m-tatar
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - TT dataset. It achieves the following results on the evaluation set:
- Loss: 0.1959
- Wer: 0.2454
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: 7e-05
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 4000
- num_epochs: 50.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.173 | 9.66 | 4000 | 0.2920 | 0.3608 |
| 0.9433 | 19.32 | 8000 | 0.2336 | 0.3026 |
| 0.8552 | 28.99 | 12000 | 0.2221 | 0.2799 |
| 0.7863 | 38.65 | 16000 | 0.1953 | 0.2479 |
| 0.7365 | 48.31 | 20000 | 0.1968 | 0.2449 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.18.3
- Tokenizers 0.11.0