Premium character LoRAs

Original synthetic characters, bundles and AI tools are in my shop:

https://rkrra.com/civitai

Get the character LoRAs on Patreon: https://www.patreon.com/cw/RkRRa

Join the community on Discord: https://discord.gg/nB79wRNtMT

Family Guy Season 05

A style LoRA for Krea 2. Applies a consistent visual treatment without locking the subject matter.

⚠️ WARNING: faces are not perfect

This model does not always render faces perfectly. I trained it several times, but with this many LoRAs (22 seasons) I ended up giving up on getting the faces right every time. The overall style is good, but sometimes the faces come out a bit off.

I'm publishing it anyway because it is usable, it took a huge number of training hours and it would be a shame to throw it away. Just keep in mind that it is not a perfect model.

Recommended weight: 1.0 – 1.5. Adjust it to your taste and try higher weights too. Generate in wide 16:9: that is the format it works best in.

Usage

Include the trigger word fmlygyssn05 in your prompt.

Recommended weight: 1.0 – 1.5 (see the warning above).

Prompt Examples

Here are some prompt examples:

"fmlygyssn05, A small child with a bald head and red overalls sits at."

"fmlygyssn05, A large, beige shopping mall with a prominent arched entrance and red."

"fmlygyssn05, A futuristic, multi-tiered school building floats in outer space, accompanied by a."

"fmlygyssn05, A bustling airport terminal with a prominent control tower and a large."

"fmlygyssn05, A white, dog-like figure in a black tuxedo and red bow tie."

"fmlygyssn05, Three animated figures in pilot uniforms are seated in an aircraft cockpit."

"fmlygyssn05, A cluttered living room is filled with stacks of identical books and."

"fmlygyssn05, A man in a cowboy hat and a woman are seated inside."

"fmlygyssn05, A white dog in a green military uniform, wearing a cap and."

"fmlygyssn05, A uniformed officer stands beside a blue pickup truck, holding a briefcase."

"fmlygyssn05, A large, shirtless man sits on a medical examination table while a."

"fmlygyssn05, A red car is parked at a gas station pump in front."

"fmlygyssn05, A large white dog and a small child, both dressed in matching."

"fmlygyssn05, A cartoon-style bar named “The Drunken Clam” stands at night, its sign."

"fmlygyssn05, A person with blonde hair, wearing a red vest, sits on a."

"fmlygyssn05, A suburban house with light yellow siding, a red front door, and."

"fmlygyssn05, Two women stand at the center of a crowded indoor gathering, engaged."

"fmlygyssn05, A red, two-story elementary school building with white trim and multiple windows."

"fmlygyssn05, Two cartoon figures in military-style camouflage gear are crouched beneath the chassis."

"fmlygyssn05, Two men in traditional head coverings stand on a sunlit street, one."

Training

• Base model: Krea 2 (trained on Raw)
• Type: Style LoRA
• Dataset: 281 images
• Steps: 2000
• Network rank / alpha: 32 / 32
• Learning rate: 0.0002
• Optimizer: adamw8bit
• Resolutions: 512 / 1024

Generation settings

Trained on Krea 2 Raw, and transfers to Krea 2 Turbo for faster generation.

• Raw: ~30 steps, guidance 3.5
• Turbo: ~8 steps, guidance 1.0

The preview images were generated with krea2_turbo_mxfp8.safetensors, at 8 steps, guidance 1, LoRA weight 1.5, in 16:9.