---
library_name: transformers
language:
- ko
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- sikaro/Old_People_Talk_Dataset_Ko_Train
model-index:
- name: Whisper_Old_People_Talk_check
  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_Old_People_Talk_check

This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the Old_People_Talk_Dataset_Ko_Train dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1054
- Cer: 7.3499

## 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: 100
- training_steps: 600
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Cer    |
|:-------------:|:------:|:----:|:---------------:|:------:|
| 0.3369        | 0.0221 | 100  | 0.2803          | 7.3760 |
| 0.1671        | 0.0443 | 200  | 0.1688          | 5.1938 |
| 0.1268        | 0.0664 | 300  | 0.1358          | 5.9881 |
| 0.1288        | 0.0885 | 400  | 0.1224          | 6.1016 |
| 0.0951        | 0.1107 | 500  | 0.1106          | 6.3809 |
| 0.1026        | 0.1328 | 600  | 0.1054          | 7.3499 |


### Framework versions

- Transformers 4.46.0.dev0
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0