whisper-small-ko-E10_Y_freq_speed-SA
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: 0.2344
- Cer: 7.0665
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: 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: 2
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Cer |
---|---|---|---|---|
0.5919 | 0.13 | 100 | 0.3346 | 7.9417 |
0.3901 | 0.26 | 200 | 0.2585 | 6.9079 |
0.289 | 0.39 | 300 | 0.2636 | 7.5129 |
0.2718 | 0.52 | 400 | 0.2478 | 7.0841 |
0.2722 | 0.64 | 500 | 0.2503 | 6.8080 |
0.2464 | 0.77 | 600 | 0.2444 | 6.8668 |
0.2295 | 0.9 | 700 | 0.2437 | 7.1076 |
0.1521 | 1.03 | 800 | 0.2393 | 7.1311 |
0.1346 | 1.16 | 900 | 0.2363 | 6.7375 |
0.1398 | 1.29 | 1000 | 0.2349 | 7.7538 |
0.1277 | 1.42 | 1100 | 0.2376 | 7.2780 |
0.1268 | 1.55 | 1200 | 0.2340 | 7.0547 |
0.1389 | 1.68 | 1300 | 0.2333 | 6.9843 |
0.1277 | 1.81 | 1400 | 0.2346 | 7.0547 |
0.1207 | 1.93 | 1500 | 0.2344 | 7.0665 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.0
- Tokenizers 0.15.0
- Downloads last month
- 4