Deep Reinforcement Learning with IsaacSim

This repository provides resources for learning and practicing Deep Reinforcement Learning using the NVIDIA IsaacSim environment. It includes installation instructions, lecture links, and reference sources.

Installation

1. Create Conda Environment

conda create -n isaacsim python=3.11
conda activate isaacsim

2. Install IsaacSim Python Package

Install IsaacSim 5.1.0 from the NVIDIA PyPI repository:

pip install "isaacsim[all,extscache]==5.1.0" --extra-index-url https://pypi.nvidia.com

3. Install Gym

Install Gym 0.26.2 and the classic control environments:

pip install gym==0.26.2
pip install "gym[classic_control]"

---

4. Install Wandb

Install Wandb to log Training

pip install wandb

---

Lecture Link

You can find the lecture videos in the playlist below:

https://youtube.com/playlist?list=PLOYlvnpEJzrpQvlP9AqXpO5JQznzA_pjk&si=xX0VZoWrL28OiuNU

---

Train and Play Your Policy

Train

python isaac_env/env/train.py --wandb

Play

python isaac_env/env/paly.py --checkpoint=[your check point path]

Source

Oh, S. (n.d.). Deep Reinforcement Learning [PowerPoint slides]. Department of Mathematics & Department of Data Science, Korea University.

---