Usage Guide

Usage Guide#

This guide covers detailed usage patterns and customization options for BraxViewer.

Basic Usage#

The typical workflow involves running the viewer server separately from your training process:

  1. Start the viewer server in one terminal

  2. Run your training script in another terminal

  3. View the visualization in your web browser

Viewer Server#

The viewer server can be started using the WebViewer class: host: the IP to host server port: the port to host server xml_string: your robot XML num_envs: the number of parallel environments in your viewer

from braxviewer.WebViewer import WebViewer

viewer = WebViewer(
    host="127.0.0.1",
    port=8000,
    xml=xml_string,
    num_envs=8
)
viewer.run()

Training Integration#

To integrate with your training process, wrap your environment using ViewerWrapper with a BraxSender instance: host: IP of the host server you want to send to port: port of the host server you want to send to xml_string: your robot XML num_envs: the number of parallel environments in your viewer

from braxviewer.BraxSender import BraxSender
from braxviewer.wrapper import ViewerWrapper

# Create sender to communicate with viewer
sender = BraxSender(
    host="127.0.0.1",
    port=8000,
    xml=xml_string,
    num_envs=8
)
sender.start()

# Wrap your environment
env = ViewerWrapper(env=env, sender=sender)

# Your Brax training codes
...

PPO integration#

from braxviewer.brax.training.agents.ppo import train as ppo

  • When env is wrapped by ViewerWrapper, ppo.train(...) automatically uses env.render_fn and dynamically reads env.sender.rendering_enabled to toggle rendering at runtime.

  • No extra callback wiring is required. modified PPO algorithm already has callback integrated.

Make sure the IP and port are both available.

Examples#

See the Quick Start Guide for simple examples:

  • examples/brax/cartpole/: Custom Cartpole environment

  • examples/brax/brax_envs/: Built-in Brax environments

Advanced Usage#

For detailed API documentation and advanced integration patterns, see the Python API documentation.

Contents