whisper-small-ko-Yspeed_SA
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: 0.2766
- Cer: 8.3177
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 |
---|---|---|---|---|
0.5856 | 0.13 | 100 | 0.3403 | 7.9535 |
0.3843 | 0.26 | 200 | 0.2805 | 8.0827 |
0.2779 | 0.39 | 300 | 0.2829 | 8.3059 |
0.2761 | 0.52 | 400 | 0.2659 | 7.5540 |
0.2491 | 0.64 | 500 | 0.2780 | 8.0710 |
0.2363 | 0.77 | 600 | 0.2836 | 7.8536 |
0.2208 | 0.9 | 700 | 0.2818 | 7.9359 |
0.1507 | 1.03 | 800 | 0.2740 | 8.1003 |
0.1266 | 1.16 | 900 | 0.2770 | 8.3764 |
0.1381 | 1.29 | 1000 | 0.2763 | 8.3940 |
0.117 | 1.42 | 1100 | 0.2785 | 8.1473 |
0.1385 | 1.55 | 1200 | 0.2782 | 8.4528 |
0.128 | 1.68 | 1300 | 0.2763 | 8.3764 |
0.1138 | 1.81 | 1400 | 0.2767 | 8.3940 |
0.1192 | 1.93 | 1500 | 0.2766 | 8.3177 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
- Downloads last month
- 6