Trained on NAI base model, training data and its tags are from danbooru
This is a complete retraining since my first Lora had major issues. This was trained with over ~402 tagged danbooru images using base NAI instead of Anything V3. The character's outfit should trigger without any specific keywords but tags that were trained are: hat, beanie, hoodie, armbands, horseshoe_ornament, star_print, blue_pants
Other notes:
add hat to add the hat
add blue lips for the optional blue lipstick
most of the dataset uses his ASBR color pallet
This should work with most if not all N*vel ai derivatives (anything, orange mixes, most anime custom models really.)
hi
my quick first try at a LORA + a friend of mine asked if I can share it so here it is for anyone else to mess with愛.. This was trained on Anything v3 and a weight of 0.9 or lower is recommended. Was tested on Anything V4 and works.
Other notes:
