LoRA trained with 50 images.
The recommended strength is: 0.8-1.0
•This model is not capable of recreating the original characters perfectly, but it can come a little close.
•It is highly recommended to use HiRes.fix, and Adetailer for full body shots or those away from the camera.
•The recommended sampler and the one that works best is Euler a, others could generate artifacts in the images.
Description
Illustrious ver.
Trained with a larger and better dataset.
FAQ
Details
Downloads
977
Platform
CivitAI
Platform Status
Available
Created
6/15/2025
Updated
4/29/2026
Deleted
-
Trigger Words:
obata takeshi
traditional media
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.