chinhtruong/kk-kontext-lora1

Katzkin Kontext LoRA Model

This is a custom-trained LoRA adapter for FLUX.1-Kontext-dev designed for premium automotive seat transformations (fabric-to-leather upholstery conversions).

Model Details

• Base Model: black-forest-labs/FLUX.1-Kontext-dev
• Trigger Word: katzkin_upholstery
• Training Steps: 1500 steps
• Training Dataset: 60 high-resolution real-world Katzkin upholstery installations.

How to use in Diffusers

from diffusers import FluxPipeline
import torch

pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16)
pipe.load_lora_weights("chinhtruong/kk-kontext-lora1", weight_name="katzkin_kontext_lora.safetensors")
pipe.fuse_lora(lora_scale=1.0)
pipe.to("cuda")

# Run inference matching your trigger word
# prompt = "katzkin_upholstery, Seat Tones Configuration: 2-Tone Face [Wrap: BLACK, Face: RED]..."