MOSS-Image project page
A static academic project page, following the presentation structure of the MOSS-Speech homepage. HTML, CSS, and JavaScript are self-contained; no build step or external fonts are required.
From the MOSS-Image repository root:
python -m http.server 5173 --bind 0.0.0.0 --directory website
Open http://localhost:5173/ or use your environment's forwarded port URL.
All local resource paths are relative, so a proxy subdirectory works without
environment-specific routing code. Keep the trailing slash in the base URL.
Content follows the repository README, including
Base/SFT model links and comparison, Diffusers Quick Start, CFG 3, user guides,
ComfyUI Manager installation, the Prompt Enhancer skill, and the QQ community link. The two showcase overview mosaics retain the
README’s original image arrangement and proportions; the introductory teaser appears above the introductory paragraph in the hero section.
The layout diagram illustrates an input specification; the showcase figures are
copied from assets/showcase/ in the repository. Images open at full resolution.
The Live Demo button links to the Hugging Face Space.
The report is linked in the repository READMEs as
paper/MOSS_Image_Technical_Report.pdf. A copy lives at the same relative path
inside website/ so the static site can be served independently. When the report
or showcase assets change, refresh the corresponding website copies too.
Starting the local preview does not publish the site.
The official homepage is https://openmoss.github.io/MOSS-Image/.
In the OpenMOSS/MOSS-Image repository, select Settings → Pages → Source →
GitHub Actions. The Deploy homepage workflow publishes website/ when changes
to that directory or the workflow are pushed to main. It can also be started
manually from Actions. The workflow deploys only from the official repository;
development forks are skipped.