Finetuned Whisper large for darija speech translation
This model is a fine-tuned version of openai/whisper-large on the Darija-C dataset. It achieves the following results on the evaluation set:
- Loss: 0.0000
- Bleu: 0.7440
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: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Use 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: 100
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu |
|---|---|---|---|---|
| 3.145 | 0.8333 | 50 | 1.7143 | 0.0074 |
| 1.5712 | 1.6667 | 100 | 0.9313 | 0.0557 |
| 0.8924 | 2.5 | 150 | 0.3534 | 0.4306 |
| 0.3915 | 3.3333 | 200 | 0.2014 | 0.5725 |
| 0.2762 | 4.1667 | 250 | 0.0874 | 0.5841 |
| 0.1136 | 5.0 | 300 | 0.0630 | 0.6672 |
| 0.0777 | 5.8333 | 350 | 0.0868 | 0.6594 |
| 0.0749 | 6.6667 | 400 | 0.0405 | 0.7117 |
| 0.0412 | 7.5 | 450 | 0.0217 | 0.7319 |
| 0.0046 | 8.3333 | 500 | 0.0414 | 0.7320 |
| 0.0516 | 9.1667 | 550 | 0.0007 | 0.7440 |
| 0.006 | 10.0 | 600 | 0.0001 | 0.7440 |
| 0.0001 | 10.8333 | 650 | 0.0005 | 0.7440 |
| 0.0006 | 11.6667 | 700 | 0.0000 | 0.7440 |
| 0.0 | 12.5 | 750 | 0.0000 | 0.7440 |
| 0.0 | 13.3333 | 800 | 0.0000 | 0.7440 |
| 0.0 | 14.1667 | 850 | 0.0000 | 0.7440 |
| 0.0 | 15.0 | 900 | 0.0000 | 0.7440 |
| 0.0 | 15.8333 | 950 | 0.0000 | 0.7440 |
| 0.0 | 16.6667 | 1000 | 0.0000 | 0.7440 |
Framework versions
- Transformers 4.47.0
- Pytorch 2.5.1+cu121
- Datasets 2.19.2
- Tokenizers 0.21.0