README.md

---
language:
- en
base_model:
- black-forest-labs/FLUX.1-Kontext-dev
pipeline_tag: image-to-image
library_name: diffusers
tags:
- Style
- Vector
- FluxKontext
- Image-to-Image
---

# Vector Style LoRA for FLUX.1 Kontext Model
This repository provides the **Vector** style LoRA adapter for the [FLUX.1 Kontext Model](https://huggingface.co/black-forest-labs/FLUX.1-Kontext-dev).
This LoRA is part of a collection of 20+ style LoRAs trained on high-quality paired data generated by GPT-4o from the [OmniConsistency](https://huggingface.co/datasets/showlab/OmniConsistency) dataset.

![Comparison02](./style_comparison_02.jpg)
![Comparison01](./style_comparison_01.jpg)

Contributor: Tian YE & Song FEI, HKUST Guangzhou.

## Style Showcase
Here are some examples of images generated using this style LoRA:

![Vector Style Example](./example-1.png)
![Vector Style Example](./example-2.png)
![Vector Style Example](./example-3.png)
![Vector Style Example](./example-4.png)
![Vector Style Example](./example-5.png)
![Vector Style Example](./example-6.png)

## Inference Example
```python
from huggingface_hub import hf_hub_download
from diffusers import FluxKontextPipeline
from diffusers.utils import load_image
import torch

# Define the style and model details
STYLE_NAME = "Vector"
LORA_FILENAME = "Vector_lora_weights.safetensors"
REPO_ID = "Kontext-Style/Vector_lora"

# Download the LoRA weights
# Make sure you have created a folder named 'LoRAs' in your current directory
hf_hub_download(repo_id=REPO_ID, filename=LORA_FILENAME, local_dir="./LoRAs")

# Load an image
image = load_image("https://huggingface.co/datasets/black-forest-labs/kontext-bench/resolve/main/test/images/0003.jpg").resize((1024, 1024))

# Load the pipeline
pipeline = FluxKontextPipeline.from_pretrained("black-forest-labs/FLUX.1-Kontext-dev", torch_dtype=torch.bfloat16).to('cuda')

# Load and set the LoRA adapter
pipeline.load_lora_weights(f"./LoRAs/{LORA_FILENAME}", adapter_name="lora")
pipeline.set_adapters(["lora"], adapter_weights=[1])

# Run inference
prompt = f"Turn this image into the {STYLE_NAME.replace('_', ' ')} style."
result_image = pipeline(image=image, prompt=prompt, height=1024, width=1024, num_inference_steps=24).images[0]
result_image.save(f"{STYLE_NAME}.png")

print(f"Image saved as {STYLE_NAME}.png")
```

Feel free to open an issue or contact us for feedback or collaboration!