yinwin / diffsynth-studio

Enjoy the magic of Diffusion models!

DiffSynth Studio

Introduction

DiffSynth Studio is a Diffusion engine. We have restructured architectures including Text Encoder, UNet, VAE, among others, maintaining compatibility with models from the open-source community while enhancing computational performance. We provide many interesting features. Enjoy the magic of Diffusion models!

Roadmap

• Aug 29, 2023. We propose DiffSynth, a video synthesis framework.
  • Project Page.
  • The source codes are released in EasyNLP.
  • The technical report (ECML PKDD 2024) is released on arXiv.
• Oct 1, 2023. We release an early version of this project, namely FastSDXL. A try for building a diffusion engine.
  • The source codes are released on GitHub.
  • FastSDXL includes a trainable OLSS scheduler for efficiency improvement.
    • The original repo of OLSS is here.
    • The technical report (CIKM 2023) is released on arXiv.
    • A demo video is shown on Bilibili.
    • Since OLSS requires additional training, we don't implement it in this project.
• Nov 15, 2023. We propose FastBlend, a powerful video deflickering algorithm.
  • The sd-webui extension is released on GitHub.
  • Demo videos are shown on Bilibili, including three tasks.
    • Video deflickering
    • Video interpolation
    • Image-driven video rendering
  • The technical report is released on arXiv.
  • An unofficial ComfyUI extension developed by other users is released on GitHub.
• Dec 8, 2023. We decide to develop a new Project, aiming to release the potential of diffusion models, especially in video synthesis. The development of this project is started.
• Jan 29, 2024. We propose Diffutoon, a fantastic solution for toon shading.
  • Project Page.
  • The source codes are released in this project.
  • The technical report (IJCAI 2024) is released on arXiv.
• June 13, 2024. DiffSynth Studio is transfered to ModelScope. The developers have transitioned from "I" to "we". Of course, I will still participate in development and maintenance.
• June 21, 2024. We propose ExVideo, a post-tuning technique aimed at enhancing the capability of video generation models. We have extended Stable Video Diffusion to achieve the generation of long videos up to 128 frames.
  • Project Page.
  • Source code is released in this repo. See examples/ExVideo.
  • Models are released on HuggingFace and ModelScope.
  • Technical report is released on arXiv.
• Until now, DiffSynth Studio has supported the following models:
  • Stable Diffusion
  • Stable Diffusion XL
  • ControlNet
  • AnimateDiff
  • Ip-Adapter
  • ESRGAN
  • RIFE
  • Hunyuan-DiT
  • Stable Video Diffusion
  • ExVideo

Installation

Create Python environment:

conda env create -f environment.yml

We find that sometimes conda cannot install cupy correctly, please install it manually. See this document for more details.

Enter the Python environment:

conda activate DiffSynthStudio

Usage (in Python code)

The Python examples are in examples. We provide an overview here.

Long Video Synthesis

We trained an extended video synthesis model, which can generate 128 frames. examples/ExVideo

github_title.mp4

Image Synthesis

Generate high-resolution images, by breaking the limitation of diffusion models! examples/image_synthesis

512*512	1024*1024	2048*2048	4096*4096

1024*1024	2048*2048

Toon Shading

Render realistic videos in a flatten style and enable video editing features. examples/Diffutoon

Diffutoon.mp4

Diffutoon_edit.mp4

Video Stylization

Video stylization without video models. examples/diffsynth

winter_stone.mp4

Chinese Models

Use Hunyuan-DiT to generate images with Chinese prompts. We also support LoRA fine-tuning of this model. examples/hunyuan_dit

Prompt: 少女手捧鲜花，坐在公园的长椅上，夕阳的余晖洒在少女的脸庞，整个画面充满诗意的美感

1024x1024	2048x2048 (highres-fix)

Prompt: 一只小狗蹦蹦跳跳，周围是姹紫嫣红的鲜花，远处是山脉

Without LoRA	With LoRA

Usage (in WebUI)

python -m streamlit run DiffSynth_Studio.py

sdxl_turbo_ui.mp4