whisper-small-sinhala-politics-5h
This model is a fine-tuned version of janiduchamika/whisper-small-sinhala-general-185k on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1543
- Wer: 28.6521
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.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: 8
- num_epochs: 8
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1211 | 1.1887 | 32 | 0.1248 | 32.2512 |
| 0.0309 | 2.3774 | 64 | 0.1280 | 30.8398 |
| 0.0144 | 3.5660 | 96 | 0.1362 | 30.2047 |
| 0.007 | 4.7547 | 128 | 0.1437 | 29.3578 |
| 0.0027 | 5.9434 | 160 | 0.1493 | 30.3458 |
| 0.0015 | 7.1132 | 192 | 0.1543 | 28.6521 |
Framework versions
- Transformers 4.57.2
- Pytorch 2.9.0+cu126
- Datasets 4.4.1
- Tokenizers 0.22.1