Yujun-Shi/DragDiffusion

Health v2 · maintenance only

Adoption not part of the score

1259 stars · 95 forks observed · 2026-08-28

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

DragDiffusion is the official CVPR 2024 research code for interactive point-based image editing using pretrained diffusion models. It provides a UI where users drag points on an image to precisely reposition content, using LoRA training and diffusion latent optimization.

Use cases

• drag points on an image to move or reshape objects
• edit real photos with point-based diffusion editing
• edit stable diffusion generated images by dragging
• reproduce DragDiffusion research results
• run DragBench evaluation for point-based editing

When to choose

• you need precise spatial control over image edits that text prompts can't achieve
• you want a research baseline or to build on point-based diffusion editing
• you have an NVIDIA GPU and want a UI for drag-based editing

When to avoid

• you need a polished commercial image editor
• you have no GPU or want CPU-only editing
• you need text-prompt-only editing without spatial control

Facets

application · maturity maintenance

image-processing machine-learning gui artificial-intelligence image-processing computer-vision deep-learning python diffusion-models stable-diffusion image-editing point-based-editing lora research-code cvpr2024 linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Yujun-Shi/DragDiffusion")