---
library_name: diffusers
license: apache-2.0
---
<!-- <p align="center">
  <img src="https://github.com/MLP-Lab/KORMo-tutorial/blob/main/tutorial/attachment/kormo_logo.png?raw=true" style="width: 100%; max-width: 1100px;">
</p> -->

<p align="center">
  <img src="https://github.com/MLP-Lab/KORMo-tutorial/blob/main/tutorial/attachment/kormo_logo.svg?raw=true" style="width: 40%; max-width: 1100px;">
</p>


## 🚀 Update News
- **2026-03-05**: Official release of KORMo-Diffusion.
- **2026-03-02**: Official release of KORMo-VL.
- **2025-10-13**: Official release of KORMo-10B-sft.
---
## 💡 About KORMo-VL-Diffusion

**KORMo-VL** is a vision-language model developed **from scratch by the KAIST MLP Lab (https://sites.google.com/view/aailab)**, built on top of **KORMo-10B**.
The system consists of two components:

* **Vision-Language Model (VLM)**
* **Image Generation Model**

The KORMo-VL-Diffusion model, designed for image generation, was trained from scratch with a high proportion of images reflecting Korean daily environments and culture. 
<span style="color:red">Unfortunately, due to limited GPU resources during the research process, we are sharing the intermediate results of the model at this stage.</span>

---

KORMo-VL은 KAIST MLP 연구실에서 **from scratch로 개발한 시각-언어 모델**로, KORMo-10B를 기반으로 (1) 시각-언어 모델과 (2) 이미지 생성 모델로 구성되어 있습니다.

이 중 **이미지 생성을 위한 KORMo-VL-Diffusion** 모델은 한국의 생활 환경과 문화를 반영하기 위해 국내 환경 이미지를 가능한 높은 비율로 사용하여 **from scratch부터 학습된 모델**입니다. 
<span style="color:red">다만 연구 진행 중 GPU 자원을 추가로 확보하지 못해 **현재는 중간 결과물을 공유하게 되었습니다.**</span>

* **LLM:** KORMo-VL
* **Model Structure:** Qwen-Image를 구조를 참조해 재개발함 (20B 정도의 Diffusion부분을 변형해 scratch부터 학습)
* **Languages:** Korean / English
* **Training Data:** Synthetic data + public datasets (e.g., AI Hub, details to be released)

향후 해당 모델을 충분히 학습할 수 있는 환경이 마련된다면 **완성된 모델로 발전시키는 것을 목표로 하고 있습니다.**
중간 결과물 위에서 추가 튜닝이나 연구를 진행하고 싶은 분들은 **자유롭게 활용해 보시기 바랍니다.**



## 📈 T2I Performance
### English Prompt
| Prompt | Generated Image |
| :--- | :--- |
| **Prompt:** Dense forest | <img src="https://huggingface.co/KORMo-VL/KORMo-VL-Diffusion/resolve/main/example_images/Dense%20forest.webp" width="400"> |
| **Prompt:** Black pattern mug | <img src="https://huggingface.co/KORMo-VL/KORMo-VL-Diffusion/resolve/main/example_images/black%20pattern%20mug%20cpup.webp" width="400"> |

### Korean Prompt
| Prompt | Generated Image |
| :--- | :--- |
| **Prompt:** 울창한 숲 | <img src="https://huggingface.co/KORMo-VL/KORMo-VL-Diffusion/resolve/main/example_images/Dense%20forest.webp" width="400"> |
| **Prompt:** 검은 무늬의 머그컵 | <img src="https://huggingface.co/KORMo-VL/KORMo-VL-Diffusion/resolve/main/example_images/%EA%B2%80%EC%9D%80%20%EB%AC%B4%EB%8A%AC%EC%9D%98%20%EB%A8%B8%EA%B7%B8%EC%BB%B5.webp" width="400"> |



## KORMo-VL-Diffusion Demo

`prompt: 아름다운 정원의 꽃들`

<video width="640" height="360" controls>
  <source src="https://huggingface.co/KORMo-VL/KORMo-VL-Diffusion/resolve/main/kormo_diffusion_assets/kormo_t2i.mp4" type="video/mp4">
</video>


## 📦 Installation

```bash
uv pip install transformers==4.57.1 pillow torchvision diffusers
```

---
## 🚀 Inference Example
```
github repo 활용 예정
```

---


## Contact
- KyungTae Lim, Professor at KAIST. `ktlim@kaist.ac.kr`

## Contributor (https://sites.google.com/view/aailab)
- Junghun Yuk
- INho won
- HANGYEOL YOO
- Junmyeong Lee
- KyungTae Lim

## Citation

```text
@misc{KORMo,
  author = {Minjun Kim, Hyeonseok Lim, Hangyeol Yoo, Inho Won, Seungwoo Song, Minkyung Cho, Junghun Yuk, Changsu Choi, Dongjae Shin, Huije Lee, Hoyun Song, Alice Oh, and KyungTae Lim},
  title = {KORMo: Korean Open Reasoning Model for Everyone},
  year = {2025},
  publisher = {GitHub},
  journal = {Technical Report},
  paperLink = {\url{https://arxiv.org/abs/2510.09426}},
 },
}
```