whisper-small_vaani_62082_ratio_alldata_lr_1e-4_lgid_en_epochs_100_FT
This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5143
- Global Wer: 15.3270
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: 0.0001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 500
- num_epochs: 100
Training results
| Training Loss | Epoch | Step | Validation Loss | Global Wer |
|---|---|---|---|---|
| 1.1567 | 1.0 | 540 | 0.4581 | 19.9970 |
| 0.6371 | 2.0 | 1080 | 0.4372 | 21.5387 |
| 0.4040 | 3.0 | 1620 | 0.4480 | 13.6806 |
| 0.2887 | 4.0 | 2160 | 0.4610 | 13.8116 |
| 0.2153 | 5.0 | 2700 | 0.4733 | 13.7779 |
| 0.1712 | 6.0 | 3240 | 0.5143 | 15.3270 |
Framework versions
- Transformers 5.0.0.dev0
- Pytorch 2.9.0+cu126
- Datasets 3.6.0
- Tokenizers 0.22.2