README_cn.md

Majesty Diffusion 👑

来用文本生成“壮丽的”图像吧！

派生于"Princess generator"

Majesty Diffusion 是基于Diffusion Model的，文本到图像(Text-to-image)的生成工具，擅长生成视觉协调的形状。 👸

访问我们的 Majestic Guide (英文网站，建设中), 或者加入我们的英文社区 on Discord。 也可以通过 @multimodalart on Twitter 或 @Dango233 on twitter 联系到作者。
Majesty Diffusion支持保存、分享、调用设定文件，如果你有喜欢的设定，欢迎一并分享出来！

更完善的中文文档正在撰写中，中文社区也即将择日开通，尽请期待 :D

本项目分两个分支：

• Latent Majesty Diffusion
• V-Majesty Diffusion

Latent Majesty Diffusion v1.5

Formerly known as Latent Princess Generator

<----点击此处即可访问Colab

Dango233 @Dango233 and apolinario (@multimodalart)合作开发的，基于 CompVis' Latent Diffusion Model开发的生成工具。模型大，擅长小尺度（256x256~256x384）下的图像生成，非常擅长生成正确的形状。如有足够显存(16GB)，可以通过内建的Upscaling获得更高分辨率的图像。

• Dango233 做了如下变更
  • 支持CLIP模型引导，提升生成质量，支持更多风格
  • 支持Upscaling(上采样)和Scheduling(步骤编排)，允许自定义Diffusion模型的不同生成阶段
  • 更好的Cutouts，以及各超参数的随时间变化的编排
  • 直接通过Clamp_max进行梯度大小的控制，更直观
  • 梯度soft clipping等一系列为提升生成质量的hack
  • 线性可变的eta schedule
  • 支持Latent diffusion的negative prompt
  • 实现了inpainting
• apolinario (@multimodalart)
  • 整理Notebook，迁移到Colab并支持本地部署
  • 实现了设定的保存、读取功能
• 其他来自社区的贡献
  • Jack000 GLID-3 XL 的无水印Fintuned模型
  • LAION-AI 基于wikiart finetune的ongo模型，更适合生成美术风格的图像
  • dmarx Multi-Modal-Comparators 用于载入CLIP及CLIP-LIKE的模型
  • 基于open_clip，实现梯度检查点，节省显存
  • crowsonkb 的Aesthetic Model 及 LAION-AI aesthetic predictor embeddings，生成更具美感的结果

V-Majesty Diffusion v1.2

Formerly known as Princess Generator ver. Victoria

A Dango233 and apolinario (@multimodalart) Colab notebook implementing crowsonkb's V-Objective Diffusion, with the following changes:

• Added Dango233 parallel multi-model diffusion (e.g.: run cc12m_1 and yfcc_2 at the same time - with or without lerping)
• Added Dango233 cuts, augs and attributes scheduling
• Added Dango233 mag and clamp settings
• Added apolinario (@multimodalart) ETA scheduling
• Added nshepperd v-diffusion imagenet512 and danbooru models
• Added dmarx Multi-Modal-Comparators
• Added crowsonkb AVA and Simulacra bot aesthetic models
• Added LAION-AI aesthetic pre-calculated embeddings
• Added open_clip gradient checkpointing
• Added Dango233 inpainting mode
• Added apolinario (@multimodalart) "internal upscaling" (upscales the output with yfcc_2 or openimages)
• Added apolinario (@multimodalart) savable settings and setting library (including defaults, disco-diffusion-defaults default settings). Share yours with us too with a pull request!

TODO

Please feel free to help us in any of these tasks!

• Figure out better defaults and add more settings to the settings library (contribute with a PR!)
• Add all notebooks to a single pipeline where on model can be the output of the other (similar to Centipede Diffusion)
• Add all notebooks to the MindsEye UI
• Modularise everything
• Create a command line version
• Add an inpainting UI
• Improve performance, both in speed and VRAM consumption
• More technical issues will be listed on https://github.com/multimodalart/majesty-diffusion/issues

Acknowledgments

Some functions and methods are from various code masters - including but not limited to advadnoun, crowsonkb, nshepperd, russelldc, Dango233 and many others