<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
whisper-small-ko-Yfreq
This model is a fine-tuned version of openai/whisper-small on the aihub elder over 70 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3317
- Cer: 7.6366
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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.1961 | 0.3 | 100 | 0.3253 | 8.7284 |
| 0.174 | 0.59 | 200 | 0.3216 | 8.6928 |
| 0.1725 | 0.89 | 300 | 0.3266 | 9.1616 |
| 0.0665 | 1.19 | 400 | 0.3156 | 8.2775 |
| 0.0601 | 1.48 | 500 | 0.3308 | 8.5979 |
| 0.0544 | 1.78 | 600 | 0.3178 | 8.5148 |
| 0.0224 | 2.07 | 700 | 0.3184 | 8.0757 |
| 0.0199 | 2.37 | 800 | 0.3334 | 8.1766 |
| 0.0272 | 2.67 | 900 | 0.3304 | 8.4317 |
| 0.0195 | 2.96 | 1000 | 0.3270 | 7.9214 |
| 0.0075 | 3.26 | 1100 | 0.3330 | 8.0460 |
| 0.0086 | 3.56 | 1200 | 0.3304 | 7.5832 |
| 0.0094 | 3.85 | 1300 | 0.3327 | 7.5892 |
| 0.0029 | 4.15 | 1400 | 0.3285 | 7.6010 |
| 0.0027 | 4.44 | 1500 | 0.3299 | 7.4171 |
| 0.0036 | 4.74 | 1600 | 0.3317 | 7.6366 |
Framework versions
- Transformers 4.34.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
