DiffuseST: Unleashing the Capability of the Diffusion Model for Style Transfer

Ying Hu, Chenyi Zhuang, Pan Gao

I2ML, Nanjing University of Aeronautics and Astronautics

Paper

⚙️ Setup and Usage

conda create --name DiffuseST python=3.8
conda activate DiffuseST

# Install requirements
pip install -r requirements.txt

Download the pre-trained blipdiffusion.

Put the content images in images/content and the style images in images/style.

# Run DiffuseST, alpha default to 0.1
python run.py

# Perform style injection for more steps
python run.py --alpha 0.2