The implementation of the proposed framework in this paper.
The conda environment file is environment.yaml. Here we use PyTorch with CUDA 10.2, which might not work with the latest GPUs. Update to newer version of PyTorch using instructions from PyTorch Install Guide
crete new environment by
conda env create -n <ENVNAME> --file environment.yaml
Before running, you might want to edit the config in the main.py
And compile the C++ extension to speed up execution.
cd gym_uav_c_functions
./build.sh
Run the experiment by:
python main.py
or run in background with:
nohup python ./main.py > out.txt 2>&1 &