lehduong/OneDiffusion
A diffusion model that unifies image synthesis and understanding across diverse vision tasks including text-to-image generation, image editing, and novel view synthesis.
Not currently ranked — collecting fresh signals.
star history
OneDiffusion is a versatile diffusion-based model published at CVPR 2025 that supports bidirectional image synthesis and understanding across multiple tasks. It handles text-to-image generation, subject-driven generation, image editing from instructions, novel view synthesis, and single-view 3D reconstruction. The model is available on HuggingFace with pre-trained weights and a demo space.
Frequently asked
- What is lehduong/OneDiffusion?
- A diffusion model that unifies image synthesis and understanding across diverse vision tasks including text-to-image generation, image editing, and novel view synthesis.
- Is OneDiffusion open source?
- Yes — lehduong/OneDiffusion is an open-source project tracked on heatdrop.
- What language is OneDiffusion written in?
- lehduong/OneDiffusion is primarily written in Python.
- How popular is OneDiffusion?
- lehduong/OneDiffusion has 662 stars on GitHub.
- Where can I find OneDiffusion?
- lehduong/OneDiffusion is on GitHub at https://github.com/lehduong/OneDiffusion.