Supported on Python 3.8 to 3.11. Expected to run on a GPU.
git clone https://github.com/y26805/stable_diffusion_with_upscaler.git
For poetry users, run
cd stable_diffusion_with_upscaler/
poetry install
For pip users, run
cd stable_diffusion_with_upscaler/
pip install -e .
git clone https://github.com/CompVis/stable-diffusion
git clone https://github.com/CompVis/taming-transformers
git clone https://github.com/CompVis/latent-diffusion
(Optional) for poetry users, run
poetry shell
Running the follow will save a file in the outputs/
folder.
python3 stable_diffusion_with_upscaler/run.py --prompt "panda in space"
Output images have a resolution of 1024x1024 pixels.
python3 stable_diffusion_with_upscaler/run.py --prompt "panda in space" --steps 50 --seed 123 --outdir myfolder
This generates an image with the text promot panda in space
, with 50 training steps and seed set as 123. The image is saved to myfolder/
.