Trained entirely on screencaps, all below noted outfits are directly from the anime
Works with other lora
Recommended weight = 0.8 - 1
The Showcase images use weight 1
All of my Showcase images use "anime_screencap" and "anime_coloring", which (based on the dataset) heavily influences the style and colors. If those colors are too strong, both tags can just be removed.
Outfit reference images are at the bottom of the description (WIP - not always accurate, yet)
Base prompt - combine with below noted outfits or use other outfit prompts
TSSTrya, short hair, green hair, facial mark, purple eyes, hair between eyes,
Standard Outfit
TryaOutfit, green dress, puffy sleeves, long sleeves, necklace, jewelry, juliet sleeves, green skirt, white shirt, pendant,
Kimono
TryaKimono, japanese clothes, white kimono, red sash, wide sleeves, yellow flower, long sleeves, floral print, print kimono,
Any kind of feedback on my models is highly appreciated!
I love seeing the images made with my models, dont be shy, post some!
Commission a character here !!
Outfit Reference Sheets (WIP - not always accurate, yet)
Description
FAQ
Comments (2)
You release your Lora so often, it surprises me, thank you. Tell me, what method do you use to train them? Do you do this locally or through third-party services?
@RussianMan_BRUTAL I do everything locally, using a 5090 with a WSL2 Linux Setup and a few self-written python scripts/helpers for queueing/bulk training :)