Whisper Small Slovenian
This model is a fine-tuned version of openai/whisper-small on the google/fleurs sl_si dataset. It achieves the following results on the evaluation set:
- Loss: 0.8831
- Wer: 39.3763
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: 64
- 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: 200
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0054 | 22.01 | 250 | 0.8189 | 39.5761 |
| 0.0015 | 45.01 | 500 | 0.8831 | 39.3763 |
| 0.0009 | 68.0 | 750 | 0.9106 | 39.5035 |
| 0.0008 | 90.01 | 1000 | 0.9193 | 39.6549 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2