Sign In

Felsi Rollo NoobAi | 3 outfits (Mobile Suit Gundam The Witch from Mercury) | LoRA

0

Download

1 variant available

SafeTensor

109.17 MB

Verified:

Type
LoRA
Stats

32

0

Reviews
Published

May 16, 2026

Base Model

NoobAI

Hash
AutoV2
0F7B1AAC91
Trigger Words
feldef, yellow eyes, fang, hair between eyes, crossed bangs, short hair, platinum blonde hair, topknot, small breasts
felGeen, green jacket, black jacket, two-tone jacket, red square-end necktie, red trimmed jacket, red shoulder boards, red emblem, green sleeves, red two-sided sleeves, green layered sleeves, green shorts, yellow boots, grey boots, multicolored boots
felBod, orange bodysuit, emblem, side cutout, waist cutout, skin tight, covered navel, orange bike shorts, black socks, no shoes
felPsGe, grey space helmet, grey spacesuit, white trim spacesuit, grey emblem,
default creator card background decoration
Followers - 1733

1.7K

Likes - 17239

17.2K

Chameleon Badge
1778940273584_11873-20260516134643.png

Felsi Rollo from Mobile Suit Gundam The Witch from Mercury. Works best between 0.8-1.0. Trained on Chenkin Noob XL-v0.2.  

Always needed: "feldef, yellow eyes, fang, hair between eyes, crossed bangs, short hair, platinum blonde hair, topknot, small breasts"  

V1:

Outfits/notes:  

- Uniform outfit: "felGeen, green jacket, black jacket, two-tone jacket, red square-end necktie, red trimmed jacket, red shoulder boards, red emblem, green sleeves, red two-sided sleeves, green layered sleeves, green shorts, yellow boots, grey boots, multicolored boots"  

- Orange Bodysuit outfit: "felBod, orange bodysuit, emblem, side cutout, waist cutout, skin tight, covered navel, orange bike shorts, black socks, no shoes"  

- Grey Pilot Suit: "felPsGe, grey space helmet, grey spacesuit, white trim spacesuit, grey emblem"  

The little Geturk house gremlin, fun slightly easier char to bake after Suletta. This might be one of my last Noobai loras as Anima has just released the full base model v1 so will be switching to that soon. I hope you all enjoy.  

Feedback and reviews are always appreciated.  

Training data is in the Lora metadata.  

Trained on 92 images.