whisper-small-fy-NL
This model is a fine-tuned version of openai/whisper-small on the CommonVoice 11 fy-NL (West-Frisian) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5276
- Wer: 0.2919
The Wer before finetuning was 1.0622.
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: 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: 500
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0 | 0 | 1.0622 | ||
| 0.9177 | 1.0 | 211 | 0.8145 | 0.3450 |
| 0.5807 | 2.0 | 422 | 0.7113 | 0.3640 |
| 0.2884 | 3.0 | 633 | 0.5276 | 0.2919 |
Framework versions
- Transformers 4.24.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.6.1
- Tokenizers 0.13.1
- Downloads last month
- 10
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.