Paddle-quadrotor-velocity-control-rlschool

reinforcement learning using PARL and Quadrotor env from rlschool

tools used

Python: Python 3.6+

paddlepaddle

PARL

rlschools

To Install prerequisites

pip install -r requirements.txt

Train

python train.py

The program will evaluate itself and print the evaluation results in the terminal.

Restore previous model

use function restore_model in train.py for example:add line in main()

restore_model(570684)

note that if the models in QuadrotorModel is modified, you cannot use previously trained models.

results

Because time is limitted, run process are not totally finished.