wav2vec2-fleur-mms-batch6-epoch16
This model is a fine-tuned version of facebook/mms-1b-FL102 on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.9813
- Wer: 0.4603
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: 0.001
- train_batch_size: 6
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 16
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.1533 | 2.86 | 100 | 1.3072 | 0.5164 |
| 1.0225 | 5.71 | 200 | 1.0065 | 0.4782 |
| 0.943 | 8.57 | 300 | 0.9750 | 0.4774 |
| 0.8974 | 11.43 | 400 | 0.9862 | 0.4564 |
| 0.8985 | 14.29 | 500 | 0.9813 | 0.4603 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0