Style LoRA Study · Flat · Casual · AI Lab

Red Hood — LoRA Training Test

Red Hood — LoRA Training Test

Stack

Role
Concept direction, generation, selection & cleanup
Base model
Anima
LoRA
Self-trained chibi game style LoRA ↓
Tools
ComfyUI, custom LoRA pipeline (Anima base)
Year
2026

From the AI Lab

Model Card Chibi Game Style — LoRA A chibi style study analyzing the shape language, proportions and rendering of commercial game art. 60-image curated dataset, dual-epoch operation.

Overview

A style LoRA for clean chibi game art — a study analyzing the shape language, proportions and rendering characteristics of commercial chibi game art, demonstrating controlled style reproduction. Used for the Chibi Party set and the Gala Diva splash.

Card

Base modelKrea 2 (open-weights, turbo)
Dataset60 curated images, prose captions, no trigger word
NetworkLoRA dim 32 / alpha 32 · final loss 0.0914
ValidationEpoch sweep; two operating points kept — a stable epoch for fidelity, a later epoch for cuter, simpler shapes
Operating notesPrompts must stay short; the style expects chibi proportions and collapses on adult ratios

Sample results

Works produced at the LoRA’s operating points — the target game style reproduced across different characters and costumes:

Sample results

Open in LAB →