# lllyasviel/Paints-UNDO

Understand Human Behavior to Align True Needs

Repository: https://github.com/lllyasviel/Paints-UNDO
Canonical: https://ross.abutalabs.com/products/paints-undo
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-08-13T22:11:05+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 36, release rhythm 35, longevity 56
- inputs: {"age_days": 787, "days_push": 385, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4067, forks 395 (observed 2026-08-28T04:08:34.122460+00:00)

## What it is
Paints-UNDO is a family of deep learning models that take an image as input and generate the step-by-step drawing sequence (sketching, inking, coloring, etc.) that could have produced it, like pressing undo many times in painting software. It ships as a local Gradio web app requiring a high-VRAM Nvidia GPU.

## Use cases
- generate a timelapse video of how an image was drawn
- reverse an artwork into its sketch and inking stages
- study human drawing behaviors for AI art research
- create undo-style drawing process videos from finished paintings
- explore step-by-step digital painting sequences locally

## When to choose
- you have a 16-24GB VRAM Nvidia GPU and want local inference
- you need drawing-process videos or intermediate art stages from a final image
- you research artist behavior modeling or AI art alignment

## When to avoid
- you have no GPU or only 8GB VRAM
- you need fast, real-time or cloud-hosted generation
- you want a polished end-user product rather than a research preview

## Facets
- artifact type: application
- maturity: active
- function: machine-learning, deep-learning, image-processing, video-processing, llm-inference
- domain: artificial-intelligence, machine-learning, image-processing, graphics
- platform: python, windows, self-hosted
- tags: diffusion-models, gradio, drawing-process, digital-painting, image-to-video, stable-diffusion, video, linux, gpu

## Member repositories
- lllyasviel/Paints-UNDO (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:34.122460+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:23:32.738245+00:00, confidence not recorded.
  - readme: https://github.com/lllyasviel/Paints-UNDO (fetched 2026-08-28T04:08:34.122460+00:00, sha 87ab2ec41f32)
- Data as of 2026-08-30T08:39:29.467469+00:00.