whisper-small-fft-afrispeech-general
This model is a fine-tuned version of openai/whisper-small on the afrispeech-200 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5428
- Wer: 0.14
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1349
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.4272 | 1.0 | 1350 | 0.5152 | 0.151 |
| 0.286 | 2.0 | 2700 | 0.4620 | 0.142 |
| 0.1252 | 3.0 | 4050 | 0.4698 | 0.14 |
| 0.0541 | 4.0 | 5400 | 0.4920 | 0.141 |
| 0.0208 | 5.0 | 6750 | 0.5202 | 0.141 |
| 0.0076 | 6.0 | 8100 | 0.5428 | 0.14 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.10.0+cu128
- Datasets 2.19.0
- Tokenizers 0.19.1