metadata
library_name: transformers
language:
- en
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- Korean_english
metrics:
- wer
model-index:
- name: Whisper tiny Korean
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Korean English
type: Korean_english
args: 'config: default, split: test'
metrics:
- name: Wer
type: wer
value: 12.536585365853659
Whisper tiny Korean
This model is a fine-tuned version of openai/whisper-tiny on the Korean English dataset. It achieves the following results on the evaluation set:
- Loss: 0.3119
- Wer: 12.5366
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: 2
- eval_batch_size: 1
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1156 | 1.0173 | 1000 | 0.2821 | 11.6829 |
| 0.0057 | 2.0346 | 2000 | 0.2956 | 12.0976 |
| 0.0065 | 3.0519 | 3000 | 0.3061 | 12.2195 |
| 0.0016 | 4.0692 | 4000 | 0.3077 | 12.3902 |
| 0.0022 | 5.0865 | 5000 | 0.3119 | 12.5366 |
Framework versions
- Transformers 4.51.0.dev0
- Pytorch 2.6.0+cu124
- Datasets 3.4.1
- Tokenizers 0.21.1