ziqihuangg/Collaborative-Diffusion
A CVPR 2023 paper implementation providing multi-modal controlled face generation and editing using pre-trained diffusion models.
Not currently ranked — collecting fresh signals.
star history
Collaborative Diffusion enables face generation and editing using multiple modalities (e.g., text, sketch, sparse facial landmarks) as control signals. The approach uses pre-trained uni-modal diffusion models in a collaborative manner during the reverse diffusion process. The method supports both synthesizing new images from multi-modal inputs and editing real images while preserving identity characteristics.
Frequently asked
- What is ziqihuangg/Collaborative-Diffusion?
- A CVPR 2023 paper implementation providing multi-modal controlled face generation and editing using pre-trained diffusion models.
- Is Collaborative-Diffusion open source?
- Yes — ziqihuangg/Collaborative-Diffusion is an open-source project tracked on heatdrop.
- What language is Collaborative-Diffusion written in?
- ziqihuangg/Collaborative-Diffusion is primarily written in Python.
- How popular is Collaborative-Diffusion?
- ziqihuangg/Collaborative-Diffusion has 441 stars on GitHub.
- Where can I find Collaborative-Diffusion?
- ziqihuangg/Collaborative-Diffusion is on GitHub at https://github.com/ziqihuangg/Collaborative-Diffusion.