gligen/GLIGEN

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2223 stars · 167 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

GLIGEN is a research library for open-set grounded text-to-image generation, extending frozen diffusion models with spatial grounding via bounding boxes, keypoints, and reference images. It provides training and inference code plus a diffusers integration for box-conditioned generation and inpainting.

Use cases

• generate images with objects placed at specified bounding boxes
• layout-to-image generation from text prompts
• inpaint regions of an image conditioned on box and concept
• ground text-to-image models on keypoints or reference images
• combine with Grounding DINO for automatic box annotation
• research on controlled diffusion image generation

When to choose

• you need spatial control over where generated objects appear
• you want grounded generation integrated with diffusers
• you are reproducing CVPR 2023 grounded text-to-image research

When to avoid

• you only need plain text-to-image without spatial grounding
• you need actively maintained production tooling rather than research code
• you prefer ControlNet-style conditioning workflows

Facets

library · maturity maintenance

image-processing machine-learning deep-learning image-processing computer-vision artificial-intelligence python text-to-image stable-diffusion grounded-generation layout-to-image diffusion-models cvpr-2023 gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="gligen/GLIGEN")