Skip to content

AiDraw

Our mission is to build community documentation based on the principles of openness and collaboration. Anyone can contribute to the creation of this document.

Quick Introduction

Stable Diffusion is a generative AI model that can produce unique and realistic images based on textual descriptions and images. It is widely used in interface prototyping, advertising creative design, poster design, and other fields. This project mainly introduces Stable Diffusion WebUi (SDWebUi), a graphical user interface (GUI) image generation framework based on Stable Diffusion.

Other related tools, such as ComfyUI, are also important parts of this ecosystem. This document contains many technical pieces, some of which may be outdated but still provide valuable technical background.

Relationship Between Stable Diffusion and NovelAI

NovelAI once leaked its second-generation Stable Diffusion model, which was trained using the danbooru online works collection, and was later supported by Stable Diffusion WebUi. Now, most models are trained based on the leaked NAI models, but due to lack of funds and unity, the art styles are mostly greasy and similar, often disobeying commands.

Later, NovelAI collected data and developed the third-generation model based on SDXL, which features diverse and obedient styles, leading most models in the field of illustration and anime.

The leak event spurred vigorous development in the text-to-image field. Now, many companies generally use proprietary style models to produce materials for visual design and other fields.

Detailed Data

Leak Incident Part 1: Includes production models and programs, size: 53.66 GB. There are two related models: 7GB and 4GB.

Leak Incident Part 2: Includes historical test codes and models, size: 124.54 GB. The related models are the same as in Part 1.

User Type Recommended Resources
Designer/Artist @秋葉aaaki's Video Tutorials, One-Click Launcher, Blueprint Node Editor
Anime Fans NovelAI

If you are a researcher, you can continue reading the archived files below.

Document Structure

Workbench Quick Access Portal

Document Section Content
Start Here Introduction to terminology and news. New updates will be posted on the news page.
Before Starting Introduction to entry requirements and basic plans.
Installation and Configuration Guide to installing and running WebUi, and handling common errors during operation.
Drawing and Debugging Explanation of CPU and NAI Leak content, application configuration, image generation error solutions, and application usage tips.
Drawing Guide Guidelines for writing prompts and practical guides.
Model Training Methods for using and training stylized models.
References Quick links page, backing up all useful links.

Access Information

URL Description
https://draw.dianas.cyou/ Main URL
https://stable-diffusion-book.vercel.app/ Vercel Version (Not accessible from mainland China)

If you cannot access images or external resources, please try changing your network. If the source site URL returns a status code of 404/300/302/502, please submit an issue to notify us to remove this content.

Some communities

Chat

CN

EN

Forum

CN

EN

Models

https://civitai.com/