Skip to content

GLIGEN

Grounded image generation from text and boxes

Visit official site

What GLIGEN is

The model is described as an open ended image generation system that works from text prompts and bounding box conditions.

What you can do with it

  • Generate images from text and bounding boxes
  • Use planned sampling to balance quality and constraints
  • Apply image based style transfer generation

Example uses

  • Generate a cat from a phrase and a box
  • Transfer style from a car image
  • Follow a semantic segmentation map

Application scenarios

  • Conditional image generation
  • Image editing
  • Image inpainting

Pros & cons

✓ What we like

  • The modular design is described as training efficient
  • Generalization to new concepts and layouts is listed

! What to watch out for

  • Conditions such as boxes must be supplied
  • Pretrained diffusion weights are kept frozen

FAQ

What conditions are used?

The material lists text descriptions and bounding boxes.

How is the base model treated?

The material states its parameters are frozen and new data is inserted.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review