Views
No views yet
1from datasets import load_dataset
2data = load_dataset('OmniGen2/OmniContext', dataset, split='train')1{
2 "task_type": "The task type of this sample",
3 "instruction": "User instruction for in-context image generation",
4 "input_images": "A list of all input images",
5 "instruction_langugae": "The language used in the instruction",
6 "key": "A unique identifier for the data sample. The generated image should be saved as {key}.png for evaluation",
7}| Primary tasks | Task type | #Sample | Input images |
|---|---|---|---|
| SINGLE | single_character | 50 | A single character image |
| SINGLE | single_object | 50 | A single object image |
| MULTIPLE | multi_character | 50 | Multiple character images |
| MULTIPLE | multi_object | 50 | Multiple object images |
| MULTIPLE | multi_character_object | 50 | A mix of character and object images |
| SCENE | scene_character | 50 | A scene image along with character images |
| SCENE | scene_object | 50 | A scene image along with object images |
| SCENE | scene_character_object | 50 | A scene image, character images, and object images |
@article{wu2025omnigen2,
title={OmniGen2: Exploration to Advanced Multimodal Generation},
author={Chenyuan Wu and Pengfei Zheng and Ruiran Yan and Shitao Xiao and Xin Luo and Yueze Wang and Wanli Li and Xiyan Jiang and Yexin Liu and Junjie Zhou and Ze Liu and Ziyi Xia and Chaofan Li and Haoge Deng and Jiahao Wang and Kun Luo and Bo Zhang and Defu Lian and Xinlong Wang and Zhongyuan Wang and Tiejun Huang and Zheng Liu},
journal={arXiv preprint arXiv:2506.18871},
year={2025}
}