StyleDiffusion Reproduction

Recreation of the Style Diffusion research paper from ICCV 2023. Paper link.

Setup Instructions

1. Clone the repository

git clone https://github.com/Ardacandra/style_diffusion_reproduction.git
cd style_diffusion_reproduction

2. Create environment

conda create -n style_diffusion_reproduction python=3.9
conda activate style_diffusion_reproduction
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
pip install -r requirements.txt
pip install git+https://github.com/openai/CLIP.git

3. Prepare dataset

python src/prepare_dataset.py --n_content 100 --size 256

4. Prepare pre-trained model from OpenAI Guided Diffusion

git clone https://github.com/openai/guided-diffusion.git
cd guided-diffusion
pip install -e .

5. Download model checkpoints

cd models/checkpoints
wget https://openaipublic.blob.core.windows.net/diffusion/jul-2021/256x256_diffusion_uncond.pt

How to Run

1. Set up the run configurations in the configs/ folder

2. Run main.py

python main.py --config configs/default.yaml