Overview

Welcome to the 🤗 Simulate tutorials! These beginner-friendly tutorials will guide you through the fundamentals of working with 🤗 Simulate. You’ll load and prepare an environment for training with your Deep RL framework of choice. Along the way, you’ll learn how to define and load different Assets, Sensors and Actuators, watch an RL Agent interact with and see what’s inside your environment and share a dataset to the Hub.

Examples

The examples assume some basic knowledge of Python, Deep Reinforcement Learning and a Deep RL framework like [Stable Baselines 3](https://stable-baselines3.readthedocs.io/en/master/). If you're already familiar with these, feel free to check out the [quictour](./quicktour) to see what you can do with 🤗 Simulate. If you want to learn more about Deep Reinforcement Learning, check of the 🤗 [Deep RL course](https://github.com/huggingface/deep-rl-class#the-hugging-face-deep-reinforcement-learning-class-).
The examples only cover the basic skills you need to use 🤗 Simulate. There are many other useful functionalities and applications that aren't discussed here, we will expand on these in upcoming releases.

To make it easy to run the examples, most of them will download an engine build from the hub. You can change this in many examples by specifying a build_exe at runtime, e.g. if you are adding a new backend feature! Here are the three cases for build_exe at runtime:

  1. (default) --build_exe="": download appropriate executable from the hub and run. This will cache a version for you to re-run!
  2. --build_exe=/local/path/: run a local development version of an executable.
  3. --build_exe=debug: run the example in the engine editor (requires installation). This can also be triggered when build_exe is set to None in the example.

If you have any questions about 🤗 Simulate, feel free to join and ask the community on our forum.

Advanced Usage

This section is for users looking to gain speed, run more advanced simulations, or use tools that are recently added.

For example, these include how to achieve high throughput training with Map Pools, the sample-factory RL framework and running Simulate on a GCP instance.

Let’s get started! 🏁