README.md

---
library_name: transformers
language:
- ko
license: apache-2.0
base_model: morish/whisper-medium-ko-v0_1_1
tags:
- morish/kresp_speech_87_48278_150000
- morish/open_communication_109_48652_150000
- morish/senior_kspeech_107_150000
- morish/telemedicine_208_150000_200000
- morish/welfare_470_150000_200000
- whisper-2024-09-06
- generated_from_trainer
model-index:
- name: whisper-ko-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-ko-finetune

This model is a fine-tuned version of [morish/whisper-medium-ko-v0_1_1](https://huggingface.co/morish/whisper-medium-ko-v0_1_1) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0829
- Cer: 2.4784

## 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: 32
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 1
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Cer    |
|:-------------:|:------:|:----:|:---------------:|:------:|
| 0.0868        | 0.2385 | 500  | 0.0859          | 2.5607 |
| 0.0856        | 0.4770 | 1000 | 0.0843          | 2.5300 |
| 0.0875        | 0.7155 | 1500 | 0.0833          | 2.4804 |
| 0.0815        | 0.9540 | 2000 | 0.0829          | 2.4784 |


### Framework versions

- Transformers 4.44.2
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1