Codes are provided as Jupyter notebook files (.ipynb). All code has been implemented and tested using Google Colab. To exclude possible compability errors, you may run the code using Google Colab (each notebook takes only couple of minutes to run).
Additionally, we provide the video "dac_cartpole_control_with_disturbance.mp4". This a 11s video that shows the succesful control of a simulated cartpole using the deep actor critic approach. To make it more interesting, we injected a disturbance to the Cartpole at second 6, from which it recovers quickly.*
*(Though, all methods can achieve this feat; You may check for yourself using the provided code).