jeju_stt
This model is a fine-tuned version of openai/whisper-base on the jeju_audio dataset. It achieves the following results on the evaluation set:
- Loss: 0.3821
- Cer: 11.9723
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 | Cer |
|---|---|---|---|---|
| 0.3689 | 2.0 | 1000 | 0.3850 | 13.3992 |
| 0.1887 | 4.0 | 2000 | 0.3482 | 11.9225 |
| 0.1057 | 6.0 | 3000 | 0.3603 | 11.8446 |
| 0.0637 | 8.0 | 4000 | 0.3821 | 11.9723 |
Framework versions
- Transformers 4.42.0.dev0
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.19.1
- Downloads last month
- 16
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
Model tree for rlaorrn/working
Base model
openai/whisper-base