LoRa trained on 50 images of British actress Jenna Coleman at 896x896 pixel resolution.
Description
FAQ
Comments (2)
May i ask you the training images are all faces not contain full body? The results showed is amazing.
I usually train on face photographs. If I want to show the body as well I normally can inpaint them onto an image of a generic eoman. I find that training on images with the body in it normally makes the face not look all that accurate in the LoRa.