Whisper Small Cantanese
This model is a fine-tuned version of poppysmickarlili/whisper-small-cantonese_02-05-2024-1727 on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3536
- Wer: 57.3413
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- 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 |
|---|---|---|---|---|
| 0.0168 | 1.3908 | 1000 | 0.3280 | 59.7446 |
| 0.0046 | 2.7816 | 2000 | 0.3460 | 58.9185 |
| 0.0005 | 4.1725 | 3000 | 0.3504 | 57.1911 |
| 0.0002 | 5.5633 | 4000 | 0.3536 | 57.3413 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.2.0
- Datasets 2.19.1
- Tokenizers 0.19.1