wav2vec2-hindi
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: 2.8814
- Wer: 1.0
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- 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: 500
- num_epochs: 40
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 23.6834 | 6.25 | 100 | 13.5748 | 1.0 |
| 8.2358 | 12.5 | 200 | 3.9834 | 1.0 |
| 3.6953 | 18.75 | 300 | 3.7861 | 1.0 |
| 3.4186 | 25.0 | 400 | 3.8232 | 1.0 |
| 3.2462 | 31.25 | 500 | 3.4688 | 1.0 |
| 2.8108 | 37.5 | 600 | 2.8814 | 1.0 |
Framework versions
- Transformers 4.20.0.dev0
- Pytorch 1.11.0+cu113
- Datasets 2.2.3.dev0
- Tokenizers 0.12.1