<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
whisper-medium-nya
This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4295
- Wer: 25.3278
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: 2.5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3156 | 0.49 | 500 | 0.6368 | 63.7448 |
| 0.2256 | 0.99 | 1000 | 0.4869 | 32.0378 |
| 0.1352 | 1.48 | 1500 | 0.4412 | 28.3284 |
| 0.1204 | 1.97 | 2000 | 0.4019 | 28.6119 |
| 0.0543 | 2.46 | 2500 | 0.4388 | 24.3473 |
| 0.0645 | 2.96 | 3000 | 0.3973 | 27.7141 |
| 0.0247 | 3.45 | 3500 | 0.4220 | 25.3160 |
| 0.0185 | 3.94 | 4000 | 0.4241 | 25.3633 |
| 0.0056 | 4.43 | 4500 | 0.4307 | 25.4696 |
| 0.0053 | 4.93 | 5000 | 0.4295 | 25.3278 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.7.1
- Tokenizers 0.13.2
