Whisper Small Hi - Swedish
This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3260
- Wer: 19.8946
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3144 | 0.65 | 500 | 0.3244 | 24.0623 |
| 0.1321 | 1.29 | 1000 | 0.2977 | 21.5563 |
| 0.1318 | 1.94 | 1500 | 0.2788 | 20.9190 |
| 0.052 | 2.59 | 2000 | 0.2852 | 20.3329 |
| 0.0203 | 3.23 | 2500 | 0.3017 | 19.8677 |
| 0.0174 | 3.88 | 3000 | 0.3008 | 19.9941 |
| 0.0083 | 4.53 | 3500 | 0.3216 | 20.0022 |
| 0.0039 | 5.17 | 4000 | 0.3260 | 19.8946 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2
- Downloads last month
- 2