whisper-small-ko-Yspeed2
This model is a fine-tuned version of openai/whisper-small on the aihub adult speed changed dataset. It achieves the following results on the evaluation set:
- Loss: 0.2744
- Cer: 10.9845
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 |
---|---|---|---|---|
0.4529 | 0.13 | 100 | 0.3102 | 7.5952 |
0.2772 | 0.26 | 200 | 0.2821 | 9.2164 |
0.2264 | 0.39 | 300 | 0.2872 | 8.2942 |
0.2172 | 0.52 | 400 | 0.2665 | 7.7185 |
0.1956 | 0.64 | 500 | 0.2784 | 10.5263 |
0.17 | 0.77 | 600 | 0.2786 | 9.4866 |
0.1787 | 0.9 | 700 | 0.2761 | 11.3017 |
0.0818 | 1.03 | 800 | 0.2732 | 10.0270 |
0.0616 | 1.16 | 900 | 0.2732 | 9.1753 |
0.0721 | 1.29 | 1000 | 0.2763 | 10.6732 |
0.0661 | 1.42 | 1100 | 0.2798 | 15.2432 |
0.0768 | 1.55 | 1200 | 0.2730 | 13.8393 |
0.0602 | 1.68 | 1300 | 0.2714 | 11.5484 |
0.0667 | 1.81 | 1400 | 0.2730 | 13.0933 |
0.0675 | 1.93 | 1500 | 0.2744 | 10.9845 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
- Downloads last month
- 3