|
--- |
|
language: |
|
- ko |
|
license: apache-2.0 |
|
tags: |
|
- hf-asr-leaderboard |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
base_model: openai/whisper-large-v2 |
|
model-index: |
|
- name: whisper_finetune |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# whisper_finetune |
|
|
|
This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on the aihub_100000 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1966 |
|
- Cer: 5.9236 |
|
- Wer: 23.0770 |
|
|
|
## 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-06 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- 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 | Cer | Validation Loss | Wer | |
|
|:-------------:|:-----:|:----:|:------:|:---------------:|:-------:| |
|
| 0.1866 | 0.16 | 1000 | 6.0386 | 0.1963 | 23.2684 | |
|
| 0.1788 | 0.32 | 2000 | 6.0483 | 0.1979 | 23.2267 | |
|
| 0.1541 | 0.48 | 3000 | 6.0116 | 0.1929 | 23.5519 | |
|
| 0.1692 | 0.64 | 4000 | 0.1966 | 5.9236 | 23.0770 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.38.0.dev0 |
|
- Pytorch 1.14.0a0+410ce96 |
|
- Datasets 2.16.1 |
|
- Tokenizers 0.15.1 |
|
|
