Training a Unitree G1 to Walk w/ Reinforcement Learning
Summary
This video demonstrates the practical application of reinforcement learning by training a Unitree G1 humanoid robot to walk using the PPO algorithm in a simulation environment (mjlab). It is highly useful for students and educators learning AI, as it provides a concrete example of advanced machine learning concepts applied to robotics.
Description
Using mjlab and PPO to train the Unitree G1 humanoid to walk inside and outside Neural Networks from Scratch book: https://nnfs.io Channel membership: https://www.youtube.com/channel/UCfzlCWGWYyIQ0aLC5w48gBQ/join Discord: https://discord.gg/sentdex Reddit: https://www.reddit.com/r/sentdex/ Support the content: https://pythonprogramming.net/support-donate/ Twitter: https://twitter.com/sentdex Instagram: https://instagram.com/sentdex Facebook: https://www.facebook.com/pythonprogramming.net/ Twitch: https://www.twitch.tv/sentdex
More Videos
44:11Unitree G1 Security Disaster
25:21Testing VLMs and LLMs for robotics w/ the Jetson Thor devkit
28:40Reinforcement learning with Unitree G1 humanoid - Dev w/ G1 P.5
30:21A bigger brain for the Unitree G1- Dev w/ G1 Humanoid P.4
29:32Unitree G1 - Moving the arms/hands - Dev w/ G1 Humanoid P.3
40:51