Whisper Small Persian
This model is a fine-tuned version of makhataei/Whisper-Small-Common-Voice on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.7217
- Wer: 45.6220
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-07
- train_batch_size: 10
- eval_batch_size: 10
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 40
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0013 | 0.15 | 100 | 0.7201 | 45.7700 |
| 0.0015 | 0.31 | 200 | 0.7220 | 45.7908 |
| 0.0019 | 0.46 | 300 | 0.7212 | 45.6382 |
| 0.0021 | 0.61 | 400 | 0.7229 | 45.7630 |
| 0.0025 | 0.77 | 500 | 0.7226 | 45.5943 |
| 0.0068 | 0.92 | 600 | 0.7229 | 46.1213 |
| 0.0293 | 1.07 | 700 | 0.7215 | 45.7075 |
| 0.0021 | 1.23 | 800 | 0.7215 | 45.6290 |
| 0.0025 | 1.38 | 900 | 0.7217 | 45.6151 |
| 0.0023 | 1.53 | 1000 | 0.7217 | 45.6220 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.1+cu117
- Datasets 2.15.0
- Tokenizers 0.15.0