Robotic Arm Manipulation with Human Experiences & HRL - Part 9 - Training the Meta Agent

This is part 9 of a video series on solving long-horizon tasks in the Franka Kitchen gym environment with Pytorch, SAC, Human Experience Replay, and a Hierarchical Model Structure(similar to HRL). In this video, we'll be training the meta agent.

For the data set used on this project, visit https://huggingface.co/datasets/rober...

For the completed code base used in this project, visit https://github.com/bobcowher/farama-k...

For the base SAC implementation I used - github.com:pranz24/pytorch-soft-actor-critic.git

For the game controller I'm using - Logitech - F310 Gaming Pad - https://www.bestbuy.com/site/logitech...

For more information on the environment - https://robotics.farama.org/envs/fran...

To talk about these projects or get help, drop me a comment or join our discord at   / discord  .

I'm not an AI practitioner, just a fan of AI posting my own personal projects. For a great intro course to reinforcement learning, I recommend checking out Phil Tabor's Udemy courses on the topic.

Finally, if you're new to Pytorch and deep learning in general, Mike Cohen's course on deep learning is the best place I can think of to start.

https://www.udemy.com/course/deeplear...

My favorite version of Linux(and what I'm using in these videos) - https://pop.system76.com/