ali-vilab/composer

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1557 stars · 47 forks observed · 2026-08-28

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

Official implementation of Composer, a 5-billion-parameter controllable diffusion model for creative image synthesis using composable conditions like text, depth, sketch, palette, and embeddings. It supports generating, manipulating, and editing images by combining multiple control signals.

Use cases

• generate images from text and depth maps
• edit images with sketches and color palettes
• do style transfer with diffusion models
• perform pose transfer on photos
• virtual try-on image generation
• region-specific image editing
• image variations and interpolations

When to choose

• you need fine-grained, composable control over image generation beyond plain text-to-image
• you want to reproduce or build on the Composer research paper
• you need image manipulation via multiple conditioning signals like depth, sketch, and palette

When to avoid

• you need a production-ready tool with guaranteed model releases and support
• you lack GPU resources for a 5-billion-parameter diffusion model
• you just want simple text-to-image generation without extra controls

Facets

library · maturity experimental

image-processing machine-learning deep-learning artificial-intelligence image-processing computer-vision python diffusion-model text-to-image controllable-generation image-editing research-code gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ali-vilab/composer")