whisper-medium-or-exp1
This model is a fine-tuned version of kavyamanohar/AI4B-Indicwhisper-or on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1437
- Wer: 40.7851
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-07
- train_batch_size: 32
- eval_batch_size: 64
- 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: 600
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3643 | 2.0533 | 100 | 0.2478 | 41.7376 |
| 0.1421 | 5.05 | 200 | 0.1664 | 41.3913 |
| 0.1 | 8.0467 | 300 | 0.1519 | 41.1459 |
| 0.0848 | 11.0433 | 400 | 0.1463 | 40.8573 |
| 0.0763 | 14.04 | 500 | 0.1444 | 40.8140 |
| 0.0726 | 17.0367 | 600 | 0.1437 | 40.7851 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.7.0+cu126
- Datasets 3.5.1
- Tokenizers 0.21.1