/AirSim-PyTorch-Drone-DDQN-Agent

This code is the result of the collaboration of RL Turkey team.

Primary LanguagePythonMIT LicenseMIT

AirSim PyTorch Drone DDQN Agent

Req: Python 3.8, Torch 1.7, CUDA 10.1,10.2,Airsim, Unreal Engine 4.25.4

We used airsim with unreal engine as env. First you have to set up your own environment with unreal or you can download our env.

Env : https://drive.google.com/drive/folders/1uoF8lSbR1EvPa0Q3npcCKXTqC0KYZwvu?usp=sharing

It would be nice if the env file and the code were in the same path.

Run:

  1. run blocks.exe
  2. run tihs command : python run.py
  3. watch your drone

Over time, it will produce outputs as follows.

You can also monitor the results with tensorboard.

My outputs: https://drive.google.com/drive/folders/1pTo4DYTc0Leq0USs-oFipWSvFAclTsuo?usp=sharinghttps://drive.google.com/drive/folders/1sAmrvG6nGnnE565lrnKZZ4uvsbVcmIn5?usp=drive_link