Skip to content

This repository leverages the Stable Diffusion model with ControlNet to create pixelated GIFs. The pipeline is designed to generate high-quality pixel art animations using advanced diffusion techniques.

License

Notifications You must be signed in to change notification settings

SangphilPark/pixel-stable-diffusion

Repository files navigation

🧭 pixel-stable-diffusion

This repository leverages the Stable Diffusion model with ControlNet to create pixelated GIFs. The pipeline is designed to generate high-quality pixel art animations using advanced diffusion techniques.

이 레포지토리는 스테이블 디퓨전 모델과 컨트롤넷을 활용하여 픽셀 GIF를 생성합니다. 디퓨전 기술을 사용하여 고품질의 픽셀 아트 애니메이션을 생성하도록 설계된 파이프라인입니다.

  • 일관성 있는 개체를 표현하기 위해 Depth, Openpose, ip-adapter 를 모두 활용합니다.

Demo Images

Here are some examples of the pixel art animations generated using this pipeline:

Main Todo RAG CHAT


Visual Workflow

fitPose fitDepth ➡️ m3
rain_28 rain_34

This sequence illustrates the transformation from the initial input frames to the final pixel art animation:

  1. Step 1: The initial depth and pose data are captured to understand the structural details of the scene.
  2. Step 2: These details are processed to maintain consistency across frames.
  3. Step 3: The processed frames are refined further to enhance the pixel art style.
  4. Final Output: The culmination of the process results in a cohesive and smooth pixel art animation.

This setup is ideal for generating high-quality pixel art GIFs with a strong focus on control and consistency.


서비스 아키텍쳐

Architecture

About

This repository leverages the Stable Diffusion model with ControlNet to create pixelated GIFs. The pipeline is designed to generate high-quality pixel art animations using advanced diffusion techniques.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published