Whisper Tiny for Central Kurdish (Soranî) - Rizgan Gerdenzerî
This model is a fine-tuned version of openai/whisper-tiny on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3716
- Wer: 58.4901
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 |
|---|---|---|---|---|
| 0.3783 | 1.2330 | 1000 | 0.4454 | 66.8478 |
| 0.2803 | 2.4661 | 2000 | 0.3868 | 60.8122 |
| 0.2231 | 3.6991 | 3000 | 0.3738 | 58.7816 |
| 0.2016 | 4.9322 | 4000 | 0.3716 | 58.4901 |
Framework versions
- Transformers 4.43.3
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 3
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for rizgan/whisper-tiny-ckb
Base model
openai/whisper-tinyDataset used to train rizgan/whisper-tiny-ckb
Evaluation results
- Wer on Common Voice 17.0self-reported58.490