metadata
language:
- ko
license: apache-2.0
base_model: openai/whisper-base
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- AIHub
model-index:
- name: test
results: []
test
This model is a fine-tuned version of openai/whisper-base on the Voice data of foreigners speaking Korean for AI learning dataset. It achieves the following results on the evaluation set:
- Loss: 0.5120
- Cer: 22.3647
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: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.0055 | 18.87 | 1000 | 0.4457 | 22.0218 |
| 0.0009 | 37.74 | 2000 | 0.4855 | 21.6916 |
| 0.0005 | 56.6 | 3000 | 0.5046 | 20.6502 |
| 0.0004 | 75.47 | 4000 | 0.5120 | 22.3647 |
Framework versions
- Transformers 4.35.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1