PRIS-CV/DemoFusion
A latent diffusion model framework that extends Stable Diffusion with progressive upscaling to generate high-resolution images without additional training.
Not currently ranked — collecting fresh signals.
star history
DemoFusion is a research framework that democratizes high-resolution image generation by extending existing open-source latent diffusion models. It employs progressive upscaling, skip residual connections, and dilated sampling mechanisms to achieve resolution outputs beyond the native capabilities of base models. The framework requires no additional training and makes high-resolution generative AI accessible without massive capital investment.
Frequently asked
- What is PRIS-CV/DemoFusion?
- A latent diffusion model framework that extends Stable Diffusion with progressive upscaling to generate high-resolution images without additional training.
- Is DemoFusion open source?
- Yes — PRIS-CV/DemoFusion is open source, released under the MIT license.
- What language is DemoFusion written in?
- PRIS-CV/DemoFusion is primarily written in Jupyter Notebook.
- How popular is DemoFusion?
- PRIS-CV/DemoFusion has 2k stars on GitHub.
- Where can I find DemoFusion?
- PRIS-CV/DemoFusion is on GitHub at https://github.com/PRIS-CV/DemoFusion.