Disclaimer: This model was trained exclusively using the Civitai on-site training tool with its default settings. It was created primarily to test and showcase the capabilities of the Civitai trainer with 0 effort.
All training is handled by Civitai (including auto tag and captions)
What does this model do?
This is a stylistic Concept LoRA trained specifically to generate cartoonish, 3D-style latex and shiny clothing. If you want to give your characters that glossy, highly-reflective stylized look while maintaining a vibrant, animated aesthetic, this LoRA is built for it!
This model was trained entirely on a personally curated, 100% SFW dataset, focusing on the material properties of latex (shine, reflections, form-fitting shapes) applied to a 3D/CGI cartoon art style.
Model Details & Training:
Base Model: Flux.2 Klein 9B
Network Dim / Rank: 32 (Offers a great balance of capturing the concept without being overly baked)
Epochs: 10
How to use it:
Recommended Weight: Start around 0.8 to 1.0. If the latex effect is too strong or altering your character's face too much, try lowering it to 0.6.
Prompting Tips: To get the best results, include keywords in your prompt that reinforce the style, such as: "digital 3D CGI," "cartoon style," "shiny latex outfit," or "glossy material."
What is it good at?
Adding shiny, rubber/latex textures to standard clothing.
Generating 3D stylized character portraits.
Creating vibrant, colorful, and glossy character designs.
Description
{
"lr": 0.0001,
"engine": "ai-toolkit",
"epochs": 20,
"ecosystem": "flux2klein",
"keepTokens": 0,
"networkDim": 64,
"numRepeats": 10,
"resolution": 1024,
"lrScheduler": "cosine",
"minSnrGamma": null,
"noiseOffset": null,
"modelVariant": "9b",
"networkAlpha": 32,
"optimizerType": "adamw8bit",
"shuffleTokens": false,
"textEncoderLr": null,
"flipAugmentation": false,
"trainTextEncoder": false
}