EYEDOL/whisper-tiny-hausa3
This model is a fine-tuned version of EYEDOL/whisper-tiny-hausa2 on the EYEDOL/naija-voices-hausa-split_0-6 dataset. It achieves the following results on the evaluation set:
- Loss: 0.6238
- Wer Ortho: 0.5015
- Wer: 0.4349
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: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 1.2512 | 1.0 | 665 | 0.6263 | 0.5023 | 0.4362 |
| 1.1143 | 2.0 | 1330 | 0.6204 | 0.4940 | 0.4280 |
| 1.0009 | 3.0 | 1995 | 0.6238 | 0.5015 | 0.4349 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 4.8.3
- Tokenizers 0.22.2