jeromeshaiju/VANGOH-LORA

🎨 SDXL LoRA: Van Gogh Style Transformer

This LoRA was fine-tuned on the Stable Diffusion XL (SDXL) Base 1.0 model. It is designed to infuse images with the iconic Post-Impressionist aesthetic of Vincent van Gogh, specifically focusing on his signature swirling brushstrokes, vibrant impasto textures, and emotive color palettes.

🚀 Model Details

• Developer: Jerome Shaiju
• Model Type: LoRA (Low-Rank Adaptation)
• Base Model: SDXL 1.0
• Trigger Word: vangoh_style
• Primary Use Case: Image-to-Image transformations (Style Transfer)

🛠️ Usage Instructions

To get the best results, use a Style Strength (Denoising) of 0.5 - 0.7. If the strength is too high, the facial features might distort; if too low, the brushstrokes won't appear.

Python Code (Diffusers Integration)

from diffusers import AutoPipelineForImage2Image
import torch

# Load Pipeline
pipe = AutoPipelineForImage2Image.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0", 
    torch_dtype=torch.float16
).to("cuda")

# Load your LoRA
pipe.load_lora_weights("jeromeshaiju/VANGOH-LORA", weight_name="vangoh-10.safetensors")

# Run Inference
prompt = "a professional portrait in vangoh_style, cinematic, high detail, masterpiece"
image = pipe(prompt=prompt, image=input_image, strength=0.6).images[0]
image.save("result.png")