A lightweight concept LoRA for creating multi-character lineups and side-profile composition group shots.
Recommended weight: 0.7 – 1.00
Triggers: side-by-side lineup, lineup
Train on Anima Base v1.0
LoRA Easy Training Scripts - Machina Fork (https://github.com/67372a/LoRA_Easy_Training_Scripts/tree/refresh)
This is my custom version of a Concept LoRA.
Other similar LoRA use higher DIM values, which often bleeds into the art style and makes it difficult to override.
Furthermore, those versions train the Text Encoder (TE), something the Anima developers explicitly recommend against, as it risks corrupting the internal encoder structure.
My version uses a low DIM setting to keep the base style clean and easy to override.
For best results (Easy to override the art style), use a lower weight around 0.7 for maximum style flexibility.
Open Source
Dataset, config, and tutorial available on Discord
https://discord.gg/jh6a8mtsQM