A model of Din that I trained here, using images I generated on CivitAI with ArtistIA's model. This was trained more with Yodayo in mind, since that site doesn't have a Din model.
Due to the lack of space for triggers, I forgot to add in "sleeveless" and "bare shoulders" during training proper. I've put them into the trigger words post-training and it seems to work, but just know they weren't part of training.
Speaking of CivitAI image training, getting the front views was easy. The back view, however, was an utter nightmare, due to CivitAI's cowardice and sheer incompetence in how conservatively they coded the yellowgate check. No matter how hard I tried wording it and how little was in, I only got one unobstructed backview.
