metadata
language:
- ko
license: apache-2.0
base_model: openai/whisper-medium
tags:
- hf-asr-leaderboard
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper_medium
results: []
whisper_medium
This model is a fine-tuned version of openai/whisper-medium on the aihub dataset. It achieves the following results on the evaluation set:
- Cer: 15.6625
- Loss: 1.4176
- Wer: 32.4788
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-07
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Cer | Validation Loss | Wer |
|---|---|---|---|---|---|
| 1.8819 | 0.01 | 100 | 11.9999 | 1.5851 | 29.7754 |
| 1.6964 | 0.02 | 200 | 14.6066 | 1.4982 | 31.2945 |
| 1.6783 | 0.02 | 300 | 14.8315 | 1.4504 | 31.7318 |
| 1.6238 | 0.03 | 400 | 15.3631 | 1.4259 | 32.1490 |
| 1.7569 | 0.04 | 500 | 15.6625 | 1.4176 | 32.4788 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.15.0
- Tokenizers 0.15.0