wav2vec2-fleur-mms-batch6-epoch24-finetunning-3
This model is a fine-tuned version of Nachuwu/wav2vec2-fleur-mms-batch6-epoch16 on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.3000
- Wer: 0.6017
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: 24
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.4497 | 2.38 | 100 | 1.6988 | 0.7297 |
| 1.3991 | 4.76 | 200 | 1.3838 | 0.6394 |
| 1.2352 | 7.14 | 300 | 1.3424 | 0.6124 |
| 1.1687 | 9.52 | 400 | 1.3295 | 0.6138 |
| 1.1241 | 11.9 | 500 | 1.3126 | 0.6067 |
| 1.0908 | 14.29 | 600 | 1.3162 | 0.6031 |
| 1.0824 | 16.67 | 700 | 1.3059 | 0.6017 |
| 1.0512 | 19.05 | 800 | 1.3027 | 0.5989 |
| 1.0258 | 21.43 | 900 | 1.2990 | 0.6031 |
| 1.0376 | 23.81 | 1000 | 1.3000 | 0.6017 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0