<!-- 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-Ymix
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: 0.3476
- Cer: 8.7106
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 |
|---|---|---|---|---|
| 0.4035 | 0.3 | 100 | 0.3146 | 8.4080 |
| 0.3067 | 0.59 | 200 | 0.3240 | 8.9954 |
| 0.2932 | 0.89 | 300 | 0.3249 | 9.0844 |
| 0.1177 | 1.19 | 400 | 0.3198 | 9.9686 |
| 0.1149 | 1.48 | 500 | 0.3288 | 8.7284 |
| 0.0958 | 1.78 | 600 | 0.3234 | 9.0785 |
| 0.0334 | 2.07 | 700 | 0.3374 | 8.8234 |
| 0.035 | 2.37 | 800 | 0.3370 | 8.6750 |
| 0.0361 | 2.67 | 900 | 0.3397 | 8.6335 |
| 0.0301 | 2.96 | 1000 | 0.3391 | 9.3693 |
| 0.0129 | 3.26 | 1100 | 0.3421 | 8.6987 |
| 0.0104 | 3.56 | 1200 | 0.3465 | 8.9302 |
| 0.0125 | 3.85 | 1300 | 0.3446 | 9.2921 |
| 0.0057 | 4.15 | 1400 | 0.3401 | 8.7996 |
| 0.0037 | 4.44 | 1500 | 0.3466 | 8.7878 |
| 0.0044 | 4.74 | 1600 | 0.3476 | 8.7106 |
Framework versions
- Transformers 4.34.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
