This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - UR dataset. It achieves the following results on the evaluation set:
- Loss: 3.8433
- Wer: 0.9852
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: 0.0003
- train_batch_size: 64
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.468 | 166.67 | 500 | 3.0262 | 1.0035 |
| 0.0572 | 333.33 | 1000 | 3.5352 | 0.9721 |
| 0.0209 | 500.0 | 1500 | 3.7266 | 0.9834 |
| 0.0092 | 666.67 | 2000 | 3.8433 | 0.9852 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0