Sign In

CitoRooN Lordgenome v1 | Lord Genome (Tengen Toppa Gurren Lagann) | Illustrious character LoRA

0

Not loaded

Download

1 variant available

fp32 SafeTensor

CitoRooN_Lordgenome_v1.safetensors

Full precision, largest file • 325.02 MB

Verified:

Type
LoRA
Stats

18

0

Generation

Not loaded

Generation License Fee

1 / 10 images

Reviews
Published

Oct 9, 2026

Base Model

Illustrious

Training
Steps: 2,150
Epochs: 8
Usage Tips
Clip Skip: 2
Strength: 0.8
Hash
AutoV2
63998873E2
default creator card background decoration
Uploads - 3

3

Followers - 58

58

Likes - 15

15

07_11_profile_ep8_0.8_s424242_00001_.png

A character LoRA for Lord Genome (Lordgenome) from Tengen Toppa Gurren Lagann: the bald head with the cracked forehead, ringed spiral eyes, handlebar moustache and goatee, in the series’ own gritty ink style.

Use

  • Trigger: lordgenome (put it first), then solo, 1boy, male focus and your scene.

  • Strength: 0.8 is the sweet spot (0.6-1.0). At 1.0 the final file can get rough on some seeds; 1.2 breaks.

  • Settings I used: Illustrious / Nova Anime XL IL v19, CLIP skip 2, DPM++ 2M Karras, 28 steps, CFG 5, 1216x832 or 832x1216.

  • Optional: the Gurren Lagann (Anime) Eyecatch Styles LoRA (trigger gurlagcatch) at 0.5 sharpens the ink grit and the spiral eyes. Not required.

  • Lighting words steer the look: the series lights him green, blue or gold scene by scene, so say what you want (green lighting, blue lighting, golden lighting).

What it does well

  • The head and face: cracked-forehead star, ringed eyes, both moustache curls.

  • Close-ups, portraits, half-body, capes, thrones.

  • It does not leak: prompts without the trigger stay a normal person (regularisation images were part of training).

Limits (I'm honest)

  • Full body is weak: he is rarely drawn standing in the source, so the face gets small at that distance.

  • Extreme close-ups with green lighting can pick up a hot pink / magenta glow (the series has green + magenta glow scenes).

  • A few seeds crop the top of the head; reroll or ask for wide shot / lower the strength.

  • Trained on the TV series’ look. The movie and manga designs are not in the data.

Training

  • Base: Nova Anime XL IL v19 (Illustrious). OneTrainer, rank 32, 8 epochs, ~2150 passes, text-encoder layer -2.

  • Data: 164 curated frames of the character (hand-rated, de-duplicated, tightly cropped where other characters shared the frame) plus 127 plain regularisation images without the trigger. Captions: trigger + content tags + shot + lighting colour.

Credits

Character © Gainax / Trigger / Aniplex. This is an unofficial fan LoRA. Credit required if you share results.