wav2vec2-common_voice-lithuanian
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the COMMON_VOICE - LT dataset. It achieves the following results on the evaluation set:
- Loss: 0.5988
- Wer: 0.6546
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: 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: 15.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 2.7 | 100 | 3.5524 | 1.0 |
| No log | 5.41 | 200 | 3.0275 | 1.0 |
| No log | 8.11 | 300 | 1.8796 | 1.0003 |
| No log | 10.81 | 400 | 0.6796 | 0.7686 |
| 3.3102 | 13.51 | 500 | 0.6373 | 0.7297 |
Framework versions
- Transformers 4.19.0.dev0
- Pytorch 1.10.0+cu113
- Datasets 2.1.0
- Tokenizers 0.12.1