wav2vec2-xlsr-53-espeak-cv-ft-sah2-ntsema-colab
This model is a fine-tuned version of facebook/wav2vec2-xlsr-53-espeak-cv-ft on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.3586
- Wer: 0.3296
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: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- 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: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.4128 | 5.71 | 400 | 0.4462 | 0.5733 |
| 0.2344 | 11.43 | 800 | 0.3489 | 0.3969 |
| 0.1181 | 17.14 | 1200 | 0.3470 | 0.3602 |
| 0.0837 | 22.85 | 1600 | 0.3608 | 0.3451 |
| 0.0645 | 28.57 | 2000 | 0.3586 | 0.3296 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.6.1
- Tokenizers 0.13.2