metadata
language:
- ko
license: apache-2.0
base_model: openai/whisper-medium
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- Marcusxx/chungnamFireStation3Kfiles
metrics:
- wer
model-index:
- name: chungnam_firestation3Kfiles_WER_model
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: Marcusxx/chungnamFireStation3Kfiles
type: Marcusxx/chungnamFireStation3Kfiles
args: 'config: ko, split: valid'
metrics:
- type: wer
value: 53.20910973084886
name: Wer
chungnam_firestation3Kfiles_WER_model
This model is a fine-tuned version of openai/whisper-medium on the Marcusxx/chungnamFireStation3Kfiles dataset. It achieves the following results on the evaluation set:
- Loss: 1.0247
- Wer: 53.2091
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0822 | 6.6667 | 1000 | 0.6869 | 122.9814 |
| 0.0145 | 13.3333 | 2000 | 0.8470 | 58.6957 |
| 0.0092 | 20.0 | 3000 | 0.8744 | 52.6915 |
| 0.0041 | 26.6667 | 4000 | 0.9301 | 58.0745 |
| 0.0009 | 33.3333 | 5000 | 0.9330 | 55.1760 |
| 0.0004 | 40.0 | 6000 | 0.9455 | 52.7950 |
| 0.0001 | 46.6667 | 7000 | 0.9811 | 52.3810 |
| 0.0 | 53.3333 | 8000 | 1.0021 | 52.6915 |
| 0.0 | 60.0 | 9000 | 1.0185 | 52.8986 |
| 0.0 | 66.6667 | 10000 | 1.0247 | 53.2091 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.2.2+cu121
- Datasets 3.2.0
- Tokenizers 0.19.1