Kang1121/GITGAN

GITGAN

Installation

1. Clone the Repository:

git clone git@github.com:Kang1121/GITGAN.git
cd GITGAN

2. Set Up the Environment:

conda env create --file environment.yaml
conda activate gitgan

Data Acquisition & Setup

1. Download Datasets:
   Fetch the required datasets from this link.

2. Place in Code Directory:
   After downloading, ensure the datasets are placed directly within the GITGAN code directory.

Configuration & Usage

Customizing Settings:
Dive into the configs directory to find and modify hyperparameters and other settings to cater to your requirements.

Running the Model:

• With Distributed Data Parallel (DDP):

torchrun --nproc_per_node=NUM_GPUs main.py --config_file configs/DATASET2RUN.yaml

Replace NUM_GPUs with the number of GPUs you wish to use.

• Without DDP:

python main.py --config_file configs/DATASET2RUN.yaml

Contact Us

E-mail: charles_kang@korea.ac.kr