INSTALL.md

Installation

Requirements

• Linux
• Python 3.5+ (Say goodbye to Python2)
• PyTorch 1.1
• CUDA 9.0+
• NCCL 2+
• GCC 4.9+
• mmcv

We have tested the following versions of OS and softwares:

• OS: Ubuntu 16.04/18.04 and CentOS 7.2
• CUDA: 9.0/9.2/10.0
• NCCL: 2.1.15/2.2.13/2.3.7/2.4.2
• GCC: 4.9/5.3/5.4/7.3

Install CG-Net

a. Create a conda virtual environment and activate it. Then install Cython.

conda create -n CG-Net python=3.7 -y
source activate CG-Net

conda install cython

b. Install PyTorch stable or nightly and torchvision following the official instructions.

c. Clone the CG-Net repository.

git clone https://github.com/WeiZongqi/CG-Net.git
cd CG-Net

d. Compile cuda extensions.

./compile.sh

e. Install CG-Net (other dependencies will be installed automatically).

pip install -r requirements.txt
python setup.py develop
# or "pip install -e ."

Note:

1. It is recommended that you run the step e each time you pull some updates from github. If there are some updates of the C/CUDA codes, you also need to run step d. The git commit id will be written to the version number with step e, e.g. 0.6.0+2e7045c. The version will also be saved in trained models.

2. Following the above instructions, CG-Net is installed on dev mode, any modifications to the code will take effect without installing it again.

Install DOTA_devkit

    sudo apt-get install swig
    cd DOTA_devkit
    swig -c++ -python polyiou.i
    python setup.py build_ext --inplace

Notice

You can run python(3) setup.py develop or pip install -e . to install CG-Net if you want to make modifications to it frequently.

If there are more than one CG-Net on your machine, and you want to use them alternatively. Please insert the following code to the main file

import os.path as osp
import sys
sys.path.insert(0, osp.join(osp.dirname(osp.abspath(__file__)), '../'))

or run the following command in the terminal of corresponding folder.

export PYTHONPATH=`pwd`:$PYTHONPATH