A simple model trained on about 17 images of rough chalk and graphite sketches from various artists. Is capable of doing both color and monochrome art for a variety of subjects. I actually recommend not upscaling if you want to keep the "rough" style for the images as some upscalers will add more detail to the image. The basic trigger word is "chalsketch", though you may need to use some other style words like "chalk", "monochrome", and "rough sketch" to get it to work in certain models.
Description
Base version of the model trained on the SD1.5 base checkpoint. Should work on other chekpoints but may require a higher strength or additional keywords. Also may randomly add angel wings to some portraits, probably because I used sentence captions instead of booru captions. Otherwise it works quite well.