/nn_cbf_visualize_learned_policies

This repository complements the NN-cbf repository and has various functionalities for plotting and visualizing the learned behavior

Primary LanguagePython

Project Description

This repo contains helper functions to simulate trained policies for RL agents in OpenAI gym and to simulate the imitation husky model.

Functionalities

Can simulate Cartpole, Pendulum and in OpenAI gym for trained policies and also for trained closed loop. Functionality to plot trajectories for system in each dimensioni including for husky.

Installation

Required packages are:

  1. Python:3.6
  2. Tensorflow-gpu-1.10.0
  3. gym
  4. h5py=2.10
  5. matplotlib
  6. pandas