# MicroDuck Startup

MicroDuck Startup is a Chinese-language, executable two-day learning path for reproducing the complete Microduck policy workflow. It deliberately stops short of physical hardware: the outcome is a verified software loop from an official baseline through local training and ONNX reintegration.

## Day 1: establish the baseline

Clone the tutorial, copy `configs/experiment.env.example`, and run the host preflight. The scripts then fetch pinned revisions of `microduck_rl`, the robot runtime, MicroDuckModels, the replica project, and the Awesome directory into an ignored workspace. Verify every resolved SHA before continuing.

Build and run the browser simulator first. This exposes the relationship between velocity commands, the 61-dimensional observation, 14 servo actions, separate ONNX behavior policies, and the handwritten switching logic. Record stability, command response, fall recovery, and behavior transitions before changing a model.

## Day 2: train and export

Install the official `microduck_rl` environment and require `torch.cuda.is_available()` to pass. Run the repository tests, then use the documented `64 env x 5 iteration` smoke test. Only after that succeeds should the main `4096 env x 1000 iteration` teaching run begin. The tutorial keeps the first pass on one GPU so distributed-system failures are not mixed with RL failures.

Locate the resulting checkpoint and export it through the official exporter. Verify that the ONNX graph accepts `[1, 61]`, produces `[1, 14]`, and includes observation normalization. Replay both official and custom policies in native MuJoCo before replacing the browser policy.

## Evidence and handoff

Finish by recording pinned upstream SHAs, GPU information, parameters, and artifact hashes in the supplied manifest. An optional worktree creates a single-variable `ENABLE_VELOCITY_PUSHES` A/B comparison. The repository also includes troubleshooting, workstation handoff, one-day profiles, and Hugging Face Space staging.

Follow the [canonical repository](https://github.com/superobk/microduck-startup) for exact commands and current pinned revisions; do not substitute floating branches during the first complete run.

Clone the tutorial and its five source repositories at the revisions recorded in configs/upstreams.env.

Pin and clone upstreams

Build MicroDuckModels, exercise official ONNX behaviors, and observe the handwritten policy-switching state machine.

Run the browser baseline

Install the official environment, list tasks, run CPU tests, and confirm CUDA visibility.

Validate microduck_rl

Run 64 environments for five iterations before starting the documented 4096-environment training run.

Smoke test and train

Use the supported exporter and verify the [1,61] to [1,14] contract with normalization embedded.

Export and verify ONNX

Compare native MuJoCo inference, install the custom policy in the browser project, and perform the optional push/no-push A/B experiment.

Replay and reintegrate

A reproducible two-day path from browser simulation and CPU checks through PPO training, ONNX export, native replay, and browser reintegration.

MicroDuck Startup: Two-Day RL Workflow

# Pair a gamepad

Bonding a controller to the robot.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

Scope

Use the upstream document for complete and current details.

Canonical source

Pair a gamepad

# Architecture

How the daemons, control loop, policies and updater fit together.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

How the daemons, control loop, policies and updater fit together.

Architecture

# Design docs

robotd, updater, remote WebRTC, boot recovery, restart order and the WebRTC console.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

robotd, updater, remote WebRTC, boot recovery, restart order and the WebRTC console.

Design docs

# Policy channel design

How community policies reach robots from the Hub: slots, the origin rule, and the reasoning behind the manifest.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

How community policies reach robots from the Hub: slots, the origin rule, and the reasoning behind the manifest.

Policy channel design

# Dev board setup

Setting up a development board and pushing branches.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

Setting up a development board and pushing branches.

Dev board setup

# Policy manifest

The schema-2 manifest.json contract: what each field means, and why only a constant-command episodic policy can become a one-shot skill.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

The schema-2 manifest.json contract: what each field means, and why only a constant-command episodic policy can become a one-shot skill.

Policy manifest

# duckctl

Controlling the robot from a laptop over Bluetooth.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

Controlling the robot from a laptop over Bluetooth.

duckctl

# Contributing guide

Upstream contribution standards.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

Contributing guide

# Roadmap

Milestones M1–M9, including the Hub model channel (M8) and autonomous brain (M9).

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

Milestones M1–M9, including the Hub model channel (M8) and autonomous brain (M9).

Roadmap

# Cheat sheet

robotctl commands for day-to-day use of the robot.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

robotctl commands for day-to-day use of the robot.

Cheat sheet

# Docs index

Map of every doc in the repo.

## Canonical Source

This entry is an editorial index. Follow the upstream document for the complete instructions and current commands.

LearnBuild & Teach

Recommended Tutorials

All Resources