<!-- 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-EYfreq
This model is a fine-tuned version of openai/whisper-small on the aihub Y dialogue dataset. It achieves the following results on the evaluation set:
- Loss: 0.2161
- Cer: 5.0021
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 |
|---|---|---|---|---|
| 0.2523 | 0.3 | 100 | 0.2529 | 6.6279 |
| 0.2278 | 0.59 | 200 | 0.2426 | 6.3193 |
| 0.2145 | 0.89 | 300 | 0.2306 | 6.3312 |
| 0.085 | 1.19 | 400 | 0.2142 | 5.5480 |
| 0.0793 | 1.48 | 500 | 0.2278 | 6.1354 |
| 0.0732 | 1.78 | 600 | 0.2077 | 5.2098 |
| 0.0266 | 2.07 | 700 | 0.2103 | 5.2394 |
| 0.0297 | 2.37 | 800 | 0.2074 | 4.9249 |
| 0.0267 | 2.67 | 900 | 0.2179 | 5.3759 |
| 0.0282 | 2.96 | 1000 | 0.2098 | 5.3522 |
| 0.0123 | 3.26 | 1100 | 0.2139 | 5.0436 |
| 0.009 | 3.56 | 1200 | 0.2142 | 5.0436 |
| 0.0096 | 3.85 | 1300 | 0.2189 | 5.2691 |
| 0.0043 | 4.15 | 1400 | 0.2137 | 4.9665 |
| 0.0036 | 4.44 | 1500 | 0.2146 | 4.9783 |
| 0.003 | 4.74 | 1600 | 0.2161 | 5.0021 |
Framework versions
- Transformers 4.34.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
