Faster version of stable diffusion running on CPU.
Took 10 seconds to generate single 512x512 image on Core i7-12700(With OpenVINO).
Based on Latent Consistency Models .
- Supports 256,512,768 image sizes
- Supports Windows and Linux
- Saves images
- Settings to control,steps,guidance,seed
- Added safety checker setting
- Maximum inference steps increased to 25
- OpenVINO support
Thanks deinferno for the OpenVINO model contribution. Currently it is tested on Windows only,by default it is disabled. We found 50% speed improvement using OpenVINO. Took 10 seconds to generate single 512x512 image on Core i7-12700
- Clone/download this repo or download release.
- Double click
install.bat(It will take some time to install,depends on your internet speed)
- To start fastsdcpu double click
start.bat
Ensure that you have Python 3.8 higher installed.
-
Clone/download this repo
-
In the terminal enter into fastsdcpu directory
-
Run the following command
chmod +x install.sh./install.sh -
To start fastsdcpu run
chmod +x start.sh./start.sh