LightDiffusion-v4.2

What's Changed

• Added image generation previewing feature using TAESD
• Fixed Image resizing using pillow
• Added interrupt button

Full Changelog: V4.1...V4.2

V4.1

Features

• Original Txt2Img, Img2Img (R-ERSGAN4x+ UltimateSDUpscaling DPM++ 2M)
• Attention syntax
• Hires-Fix (euler ancestral normal)
• GPU only
• Xformers and Pytorch optimization
• Stable-Fast implementation offering a 70% speedup at the cost of pre inference model optimization windup time
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• DPM Adaptive Karras
• Clip Skip
• LoRa and textual inversion (embeddings) support
• Automatic Prompt-Enhancing with llama3.1 (ollama)
• Discord bot integration with the installation of Boubou and the usage of LightDiffusion's pipeline
• Automatic Face and body Detailer (based on Impact pack)

Installation

To install, please download a release or clone this repo and execute the run.bat file in a cmd and you should be good to go.

From Source

Else install the python dependencies by writing pip install -r requirements.txt

After doing that, add your SD1/1.5 safetensors model to the checkpoints directory and you should be good to go.

Stable-Fast

To use the stable-fast optimization refer to this guide to install from the wheel file.

Ollama

To use the Prompt enhancer refer to this guide to install and run those commands
ollama run llama3.1
pip install ollama

Discord

To use LightDiffusion in discord, refer to the installation guide in this repo

What's Changed

• Auto detailer by @Aatrick in #5

Full Changelog: V4.0...V4.1

Flux-V1.1

LightDiffusion - Flux.Dev version 1.1

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the
least number of lines and in the least complex way. It's made by retro-engineering Stable-Diffusion, ComfyUI and
SDWebUI.

Features

• Original Txt2Img euler simple
• Attention syntax
• Xformers and Pytorch optimization
• Q8 Quantized Flux Dev model
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• textual inversion (embeddings) support
• Automatic Prompt-Enhancing with llama3.2 (ollama)

Installation

To install, please clone this repo and execute in cmd the run.bat file and you should be good to go. Be aware that you need at least 25GB of free space on your hard drive to run this program (40-50GB is recommended to offload the model out of RAM and into the drive).

From Source

Else install the python dependencies by writing pip install -r requirements.txt

After doing that, add your Q8 Flux Dev model to the Unet directory and you should be good to go.

Ollama

To use the Prompt enhancer refer to this guide to install and run those commands
ollama run llama3.2
pip install ollama

Tips and Tricks

Be aware that the prompt enhancer is not perfect and might not work as expected. If you have any issues, uncheck and rechek the prompt enhancer checkbox in the GUI.

Flux works best with resolutions corresponding to megapixel sizes, refer to this guide for more information.

Full Changelog: Flux-V1.0...Flux-V1.1

Flux-V1.0

LightDiffusion - Flux.Dev version 1.0

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the
least number of lines and in the least complex way. It's made by retro-engineering Stable-Diffusion, ComfyUI and
SDWebUI.

Features

• Original Txt2Img euler simple
• Attention syntax
• Xformers and Pytorch optimization
• Q8 Quantized Flux Dev model
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• textual inversion (embeddings) support
• Automatic Prompt-Enhancing with llama3.2 (ollama)

Installation

To install, please clone this repo and execute the run.bat file and you should be good to go.

From Source

Else install the python dependencies by writing pip install -r requirements.txt

After doing that, add your Q8 Flux Dev model to the Unet directory and you should be good to go.

Ollama

To use the Prompt enhancer refer to this guide to install and run those commands
ollama run llama3.2
pip install ollama

Full Changelog: V4.0...Flux-V1.0

LightDiffusionV4.0

LightDiffusion

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the
least number of lines and in the least complex way. It's made by retro-engineering Stable-Diffusion, ComfyUI and
SDWebUI.

Features

• Original Txt2Img, Img2Img (R-ERSGAN4x+ UltimateSDUpscaling DPM++ 2M)
• Attention syntax
• Hires-Fix (euler ancestral normal)
• GPU only
• Xformers and Pytorch optimization
• Stable-Fast implementation offering a 70% speedup at the cost of pre inference model optimization windup time
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• DPM Adaptive Karras
• Clip Skip
• LoRa and textual inversion (embeddings) support
• Automatic Prompt-Enhancing with llama3.1 (ollama)
• Discord bot integration with the installation of Boubou and the usage of LightDiffusion's pipeline
• Automatic Face and body Detailer (based on Impact pack)

Installation

To install, please place your SD1.5 model in _internal/checkpoints and execute the run.bat file and you should be good to go.

What's Changed

• Auto detailer by @Aatrick in #4

Full Changelog: V3.0...V4.0

LightDiffusionV3.0

LightDiffusion

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the
least number of lines and in the least complex way. It's made by retro-engineering Stable-Diffusion, ComfyUI and
SDWebUI.

Features

• Original Txt2Img, Img2Img (R-ERSGAN4x+ UltimateSDUpscaling DPM++ 2M)
• One click install and run (once you've added your .safetensors model)
• Attention syntax
• Hires-Fix (euler ancestral normal)
• GPU only
• Xformers and Pytorch optimization
• Stable-Fast implementation offering a 70% speedup at the cost of pre inference model optimization windup time
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• DPM Adaptive Karras
• Clip Skip
• LoRa and textual inversion (embeddings) support

Full Changelog: V1.0...V3.0

LightDiffusion V2.0

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the
least number of lines and in the least complex way. It's made by retro-engineering Stable-Diffusion, ComfyUI and
SDWebUI.

Features

• Original Txt2Img, Img2Img (R-ERSGAN4x+ UltimateSDUpscaling DPM++ 2M)
• One click install and run (once you've added your .safetensors model)
• Attention syntax
• Hires-Fix (euler ancestral normal)
• GPU only
• Xformers and Pytorch optimization
• FP16 and FP32 precision support
• Saved state in between starts
• GUI
• DPM Adaptive Karras
• Clip Skip
• LoRa and textual inversion (embeddings) support

LightDiffusion V1.0

The purpose of this repository and project is to remake the famous stable-diffusion in only one python script, with the least number of lines and in the least complex way.

Features

• Original Txt2Img
• One click install and run (once you've added your .safetensors model)
• Attention syntax
• NVIDIA GPU only (with xformers)
• Saved state in between starts
• GUI
• DPM Adaptive

LightDiffusion V0

Full pre-release stable-diffusion solo script with :

• SD1.5, SD2.0, SDXL, SD Cascade support
• xformers and pytorch optimizations
• UltimateSDUpscale
• DPM Adaptive support
• LoRa and Controlnet

Full Changelog: https://github.com/Aatrick/LightDiffusion/commits/V0