Skywork UniPic: Unified Autoregressive Modeling for Visual Understanding and Generation
Peiyu Wang, Yi Peng, Yimeng Gan, Liang Hu, Tianyidan Xie, Xiaokun Wang, Yichen Wei, Chuanxin Tang, Bo Zhu, Changshi Li, Hongyang Wei, Eric Li, Xuchen Song, Yang Liu, Yahui Zhou
2025-08-06
Summary
This paper talks about Skywork UniPic, a large AI model that can understand images, generate images from text, and edit images all in one system, working efficiently even on normal computers.
What's the problem?
The problem is that many AI models focus on only one task like understanding images or creating images, and big models usually need very powerful and expensive computers to run.
What's the solution?
Skywork UniPic solves this by combining multiple tasks into a single autoregressive model of 1.5 billion parameters that can perform image understanding, text-based image generation, and image editing while still running well on common, less powerful hardware.
Why it matters?
This matters because it makes advanced visual AI technology more accessible to people without supercomputers, allowing wider use in areas like graphic design, photo editing, and multimedia creation.
Abstract
Skywork UniPic, a 1.5 billion-parameter autoregressive model, unifies image understanding, text-to-image generation, and image editing with state-of-the-art performance on commodity hardware.