whisper-small-pashto-stage2-fleursonly
This model is a fine-tuned version of Adnan666/whisper-small-pashto-run10-augfromscratch on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6079
- Wer: 42.9132
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: 5e-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 471
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2798 | 0.1587 | 25 | 0.6543 | 43.3987 |
| 0.2595 | 0.3175 | 50 | 0.6287 | 43.1186 |
| 0.2604 | 0.4762 | 75 | 0.6143 | 43.5107 |
| 0.2326 | 0.6349 | 100 | 0.6048 | 43.6041 |
| 0.2391 | 0.7937 | 125 | 0.6025 | 42.5210 |
| 0.2705 | 0.9524 | 150 | 0.6035 | 42.9692 |
| 0.1907 | 1.1111 | 175 | 0.6052 | 42.4837 |
| 0.1899 | 1.2698 | 200 | 0.6092 | 43.1559 |
| 0.1686 | 1.4286 | 225 | 0.6091 | 42.7264 |
| 0.2113 | 1.5873 | 250 | 0.6079 | 42.9132 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.0+cu118
- Datasets 2.19.0
- Tokenizers 0.19.1