Installation

We have tested the following versions of OS and softwares:

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

conda create -n AerialDetection python=3.7 -y
source activate AerialDetection

conda install cython

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

c. Clone the AerialDetection repository.

git clone https://github.com/dingjiansw101/AerialDetection.git
cd AerialDetection

d. Compile cuda extensions.

./compile.sh

e. Install AerialDetection (other dependencies will be installed automatically).

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

Note:

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

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

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

If there are more than one AerialDetection 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

Provide feedback