Stability AI lately launched its first official model of Stable Diffusion XL (SDXL) v1.0. If you’re unfamiliar with Stable Diffusion, right here’s a quick overview:
Stable Diffusion is a text-to-image AI mannequin developed by the startup Stability AI. It takes a immediate and generates pictures based mostly on that description. What units Stable Diffusion other than different in style AI picture fashions like OpenAI’s Dall-E2 or MidJourney is that it’s open supply. This signifies that anybody can use it or contribute to its growth.
In this text, I’ll present you a step-by-step information on the way to arrange and run the SDXL 1.0 mannequin with Automatic1111’s WebUI.
My System Requirements
I’m operating on a Windows PC, and that is my present setup
- Platform: Windows PC
- Operating System: Windows 11 64-bit
- Graphics card: Nvidia RTX3060 Ti
- VRAM: 8 GB
- RAM: 16 GB
For Apple M1 chip customers, it would additionally work, however the efficiency could also be sluggish.
Step #1: Get the mission from GitHub
Check out the GitHub mission in your native machine. Here are the recordsdata you’re anticipated to see.
Step #2: Download the checkpoint and weights
The weights for SDXL can be found on HuggingFace. Download each recordsdata and place them within the stable-diffusion-webuimodelsStable-diffusion listing.
This is what your Stable-diffusion folder ought to seem like:
