Sign In

[IL] Jimmy Neutron Style / ジミー・ニュートロンの作風

69

243

576

19

Updated: Jul 7, 2025

style3dcgi2000sjimmy neutron

Verified:

SafeTensor

Type

LoRA

Stats

243

576

279

Reviews

Published

Jul 7, 2025

Base Model

Illustrious

Training

Steps: 2,786
Epochs: 14

Usage Tips

Clip Skip: 1

Trigger Words

a_j_n_s
3d

Hash

AutoV2
144BFCD46B

Created on Civitai

The Adventures of Jimmy Neutron, Boy Genius was a CGI cartoon that ran between 2002 and 2006, serving as a continuation of the 2001 movie Jimmy Neutron: Boy Genius. While not the first CGI cartoon, its use of commercially available “off the shelf” development tools and rejection of then standard for medium pseudo-realistic (often using motion capture and face scans) stylings for its humans in favor of more traditional cartoon designs (even implementing traditional animation techniques like "squash and stretch" in some action shots) represented a clear generational divide from earlier CGI animation that used in-house, proprietary tools to render realistically proportioned figures (and either avoided focusing on human(oid) characters, look horribly dated in retrospect, and/or was so outrageously expensive it near singlehandedly brought financial ruin upon a previously wildly successful cultural juggernaut).

Checkpoint: Need something suitable for 3D. I know Hassaku XL (1.3 Style A speciffically?) and "NTR Mix V777 (For LoRA)" both work OK. I welcome any other suggestions. Per the recommendation of schmede, the checkpoint Gehena works well
Tags: A_J_N_S, 3D
Negative: Cel-shading, outline (sometimes needed)
Hands: It's recommended to give the hands something to do, even just "arms at sides", as hands are prone to degeneration if left unspecified.
Should be compatible with most character LoRAs, but struggles with those trained heavily on 3D models (including the ones I've made for this series) since they already have an idea of what "3D" is supposed to look like. Some characters with very strong inherent style (100+ images with consistent style in training data) may also produce weird effects. Playing with weights may help either category.

When I made my Cindy LoRA, I noticed it actually had a nice style in 3D, even if not necessarily a perfect match for the show (though I included at least two different art styles in the training data). Since I had the DVD set I decided I’d make a bunch of Jimmy Neutron LoRAs to work towards building a style dataset. I took portions of those (purging similar looking images and very repeated characters), striped the non-series data (stuff from the movie, books, Fairly Oddparents style), and removed all the character trigger. To that I added a collection of 34 landscape, crowd, tertiary character and object focused shots I collected while getting the components to round out the data set. At 398, it will be the second largest I worked on (Gundam X style still tops it) and it will function as an endpoint for my half year+ of daily LoRAs (though it won't be the actual last LoRA at all: I've got a bunch of Pony conversions to post before I'm missing a day, and there's a few partially done things I've got laying around I may finish). Cost me 957 buzz, so please post some images.

I welcome any recommendations for good checkpoints for polygonal images. My best known are Hassaku and "NTR Mix V777 (For LoRA)". Be sure to check my profile with X and XXX not visible to see ALL the LoRAs I've made (including many Jimmy Neutron characters), as they won't show up otherwise.