Whisper-FineTuned-DL-Twi
This model is a fine-tuned version of openai/whisper-tiny on the Twi- native Ghanaian language. dataset. It achieves the following results on the evaluation set:
- Loss: 0.0063
- Wer: 17.9909
- Cer: 17.8277
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: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.0201 | 0.6333 | 1000 | 0.0259 | 40.8149 | 33.1641 |
| 0.006 | 1.2666 | 2000 | 0.0104 | 21.3751 | 15.7877 |
| 0.009 | 1.8999 | 3000 | 0.0070 | 17.9131 | 16.0698 |
| 0.0034 | 2.5332 | 4000 | 0.0063 | 17.9909 | 17.8277 |
Framework versions
- Transformers 4.48.0.dev0
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0