SkyworkUniPic is an open source multimodal model developed by SkyworkAI that focuses on three core functions: image understanding, text-generated images and image editing. It integrates these visual language tasks using a single 150 million parameter architecture, allowing developers to handle multiple image-related tasks under a unified framework. The model performs well in benchmarks such as GenEval and DPG-Bench, proving its capability in image generation and understanding.
The model is under MIT license, and both the code and model weights are open on GitHub, encouraging developers to use and modify them freely. This open source strategy significantly lowers the threshold for developers to explore visual AI applications, while also facilitating community contributions.
This answer comes from the articleSkyworkUniPic: An Open Source Model for Unified Processing Image Understanding and GenerationThe
