whisper-small-ko-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: 0.2816
- Cer: 8.0005
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 |
---|---|---|---|---|
0.6525 | 0.13 | 100 | 0.3531 | 8.5938 |
0.4301 | 0.26 | 200 | 0.2857 | 8.5879 |
0.3031 | 0.39 | 300 | 0.3006 | 8.6760 |
0.3082 | 0.52 | 400 | 0.2781 | 7.8947 |
0.29 | 0.64 | 500 | 0.2854 | 8.2237 |
0.2689 | 0.77 | 600 | 0.2839 | 8.1180 |
0.2532 | 0.9 | 700 | 0.2868 | 8.2061 |
0.1786 | 1.03 | 800 | 0.2843 | 8.6231 |
0.1446 | 1.16 | 900 | 0.2839 | 8.3882 |
0.157 | 1.29 | 1000 | 0.2855 | 8.2413 |
0.1468 | 1.42 | 1100 | 0.2828 | 7.9770 |
0.1503 | 1.55 | 1200 | 0.2796 | 8.0710 |
0.1464 | 1.68 | 1300 | 0.2800 | 7.9711 |
0.1377 | 1.81 | 1400 | 0.2819 | 8.1356 |
0.1329 | 1.93 | 1500 | 0.2816 | 8.0005 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
- Downloads last month
- 5