ChenWu98/cycle-diffusion
PyTorch implementation of CycleDiffusion, a unified latent space framework for stochastic diffusion models enabling zero-shot image-to-image translation and editing.
Not currently ranked — collecting fresh signals.
star history
This repository provides the official implementation of a paper that formalizes the ‘random seed’ concept in diffusion models and proposes methods to infer it from image pairs. The work enables zero-shot image editing and guidance by exploiting the latent space structure of stochastic diffusion models. It integrates with HuggingFace Diffusers and includes a demo on HuggingFace Spaces.
Frequently asked
- What is ChenWu98/cycle-diffusion?
- PyTorch implementation of CycleDiffusion, a unified latent space framework for stochastic diffusion models enabling zero-shot image-to-image translation and editing.
- Is cycle-diffusion open source?
- Yes — ChenWu98/cycle-diffusion is an open-source project tracked on heatdrop.
- What language is cycle-diffusion written in?
- ChenWu98/cycle-diffusion is primarily written in Python.
- How popular is cycle-diffusion?
- ChenWu98/cycle-diffusion has 659 stars on GitHub.
- Where can I find cycle-diffusion?
- ChenWu98/cycle-diffusion is on GitHub at https://github.com/ChenWu98/cycle-diffusion.