Whisper Base Shona - S2S-M Curriculum 3000 steps
This model is a fine-tuned version of openai/whisper-base on the Cleaned Google WAXAL Shona dataset. It achieves the following results on the evaluation set:
- Loss: 0.4319
- Wer: 38.3891
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_fused 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: 200
- training_steps: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.5823 | 0.7194 | 600 | 0.5640 | 48.7739 |
| 0.3761 | 1.4388 | 1200 | 0.4798 | 41.6642 |
| 0.2204 | 2.1583 | 1800 | 0.4506 | 39.9044 |
| 0.4718 | 2.8777 | 2400 | 0.4322 | 39.9642 |
| 0.2857 | 3.5971 | 3000 | 0.4319 | 38.3891 |
Framework versions
- Transformers 5.14.1
- Pytorch 2.12.0+cu130
- Datasets 5.0.0
- Tokenizers 0.22.2