Flim13

style lora for the 3D animator Flim13, though really their style is pretty simple 3D, the way they do expressions in their animation is really good even on still frames and i wanted to capture that.

used about 900 images for the dataset, mostly screenshots of their videos.

i think its managed to capture the facial expressions really well, though its also captured the low poly 3D models as well lol, but that can be tweaked by having the Lora at aound 0.5-0.8 from my testing if i wanted to get rid of most of the low polygon 3D model showing up but still keep the facial expressions of the style showing.

biggest thing holding me back from fully replicating Flim13's expressions are really just my lack of being able to come up with words to describe them lol. their smirk's and smug faces are my go to since thats the most standout and main reason i made this Lora.

waiIllustriousSDXL_v150 was used as the training model for this as thats usually my default, but it should work on most other anime style Models.

i have noticed though, if you used a character Lora that is already in a 3D style, then even at "1" strength this style gets taken over by the character Lora. in my examples im pretty sure i used a mix of Character Lora and just pulling from the Models dataset, but i don't think i had to lower the strength of the character lora.