Spanning the Visual Analogy Space with a Weight Basis of LoRAs
Paper • 2602.15727 • Published • 13
How to use gokaygokay/Low-Poly-Kontext-Dev-LoRA with Diffusers:
pip install -U diffusers transformers accelerate
import torch
from diffusers import DiffusionPipeline
from diffusers.utils import load_image
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-Kontext-dev", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("gokaygokay/Low-Poly-Kontext-Dev-LoRA")
prompt = "Convert this image to low poly version"
input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png")
image = pipe(image=input_image, prompt=prompt).images[0]


Convert this image to low poly version
Weights for this model are available in Safetensors format.
Training was done using fal.ai/models/fal-ai/flux-kontext-trainer/playground.
This adapter is one of the Hugging Face LoRAs used to construct the custom inference dataset in Spanning the Visual Analogy Space with a Weight Basis of LoRAs (LoRWeB), accepted at ECCV 2026.
If this model is useful in your research, please cite both the model and the paper that documents its academic use:
@software{aydogan2025low_poly_kontext,
author = {Aydoğan, Gökay},
title = {Low Poly Kontext Dev LoRA},
year = {2025},
url = {https://huggingface.co/gokaygokay/Low-Poly-Kontext-Dev-LoRA},
note = {Hugging Face model}
}
@article{manor2026spanning,
author = {Manor, Hila and Gal, Rinon and Maron, Haggai and Michaeli, Tomer and Chechik, Gal},
title = {Spanning the Visual Analogy Space with a Weight Basis of LoRAs},
journal = {arXiv preprint arXiv:2602.15727},
year = {2026},
doi = {10.48550/arXiv.2602.15727}
}
Maintainer: Gökay Aydoğan
Base model
black-forest-labs/FLUX.1-Kontext-dev