Wangt-CN/DisCo

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1072 stars · 101 forks observed · 2026-08-28

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

DisCo is a CVPR 2024 research codebase for referring human dance generation, producing realistic dance images and videos from a reference human subject, background, and target pose using disentangled diffusion-based control. It includes model training/inference code, a temporal module for video synthesis, and demo notebooks.

Use cases

• generate dance videos from a reference person image and pose sequence
• transfer a human pose onto an unseen subject while preserving appearance
• compose a person, background, and pose from different source images
• animate a character following skeleton keypoints
• research on controllable human image and video synthesis

When to choose

• you need pose-conditioned human image/video generation with faithful foreground and background preservation
• you want a research baseline or demo for dance synthesis
• you need compositional control over subject, background, and pose from separate sources

When to avoid

• you need a production-ready, supported product rather than research code
• you lack GPU resources for diffusion model inference or training
• your task is generic text-to-image generation without pose control

Facets

library · maturity maintenance

image-processing video-processing machine-learning deep-learning computer-vision image-processing artificial-intelligence deep-learning python diffusion-models pose-transfer dance-generation controlnet aigc cvpr2024 human-synthesis video linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Wangt-CN/DisCo")