First install WSL2 as suggested by the many different websites out there. I installed it like this:<\/p>\n\n\n\n
wsl --install -d Ubuntu-22.04<\/code><\/pre>\n\n\n\nOpen the installed Ubuntu and install the necessary packages for stable diffusion itself:<\/p>\n\n\n\n
sudo apt update && sudo apt -y upgrade && sudo apt -y install git-lfs python3-pip\npip install torch --extra-index-url https:\/\/download.pytorch.org\/whl\/cu117\npip install diffusers transformers==4.26 scipy ftfy accelerate<\/code><\/pre>\n\n\n\nAdd CUDA-Support for the WSL (from this site<\/a>):<\/p>\n\n\n\nwget https:\/\/developer.download.nvidia.com\/compute\/cuda\/repos\/ubuntu2204\/x86_64\/cuda-keyring_1.0-1_all.deb\nsudo dpkg -i cuda-keyring_1.0-1_all.deb\nsudo apt-get update\nsudo apt-get -y install cuda<\/code><\/pre>\n\n\n\nWrite yourself a small python file (e.g. as main.py) which will enable the image generation and save the images to a file with the prompt name:<\/p>\n\n\n\n
import torch\nfrom torch import autocast\nfrom diffusers import StableDiffusionPipeline\n\npipe = StableDiffusionPipeline.from_pretrained(\"CompVis\/stable-diffusion-v1-4\")\npipe.to(\"cuda\")\n\nprompt = \"a dark environment, two warriors standing on a chess field with swords drawn\"\nsteps = 50\nwidth = 512\nheight = 512\n\nwith autocast(\"cuda\"):\n for i in range(4):\n output = pipe(prompt, width=width, height=height, num_inference_steps=steps)\n image = output[\"images\"][0]\n file = prompt.replace(\" \", \"_\").replace(\",\", \"\")\n image.save(f\"{file}-{i}.png\")\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"First install WSL2 as suggested by the many different websites out there. I installed it like this: Open the installed Ubuntu and install the necessary packages for stable diffusion itself: Add CUDA-Support for the WSL (from this site): Write yourself a small python file (e.g. as main.py) which will enable the image generation and save … Stable Diffusion on WSL on Windows<\/span> weiterlesen