<!-- 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-small-ko-general-1159h
This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1691
- Wer: 21.4556
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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- training_steps: 12322
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.0937 | 0.33 | 2053 | 0.2091 | 29.9076 |
0.0788 | 0.67 | 4106 | 0.1894 | 24.9382 |
0.073 | 1.0 | 6159 | 0.1779 | 24.5723 |
0.0499 | 1.33 | 8212 | 0.1759 | 30.7157 |
0.051 | 1.67 | 10265 | 0.1711 | 25.0770 |
0.0479 | 2.0 | 12318 | 0.1691 | 21.4556 |
Framework versions
- Transformers 4.28.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.11.0
- Tokenizers 0.13.2