Simple Stable Diffusion script

Purpose of this project

Stable diffusion offline mode running script.

Directory structure

+- checkpoints
|   +- configs
|   +- dreamshaper-8
|   +- dreamshaper-xl
|   +- stable-diffusion-v1.5
|   +- stable-diffusion-v2.1
|   +- sdxl-turbo
|   +- upscaler
|       +- Real-ESRGAN
+- configs
|   +- diffuserbm.yaml
+- README.md
+- main.py
+- requirements.txt
+- requirements-maxos.txt
+- requirements-windows.txt

Making each directory

checkpoints

Download safetensors file from huggingface or the other repositories.

configs

Not yet ready to explain

sdxl-turbo

Clone from huggingface without git-lfs. In this project, use only configuration files so there is no need download other model checkpoints.

upscaler

Not yet ready to explain

Performance Check

In my RTX-3070 GPU + Ryzen 5 5600x 6-Core process, this code shows loading SDXL Turbo model in 20sec. I'll record the loading and inference speed on the other system A100 and H100.