blackotters/sonicstar

★ 17PythonAudience · researcherComplexity · 5/5Setup · hard

Mindmap

mindmap
  root((sonicstar))
    What it does
      Text-to-robot-motion
      VLA inference
      Autonomous tasks
    Training pipeline
      Collect demonstrations
      Build dataset
      Train SonicLatent
    Deployment
      Policy server
      Simulation env
      Real robot future
    Repo structure
      starVLA training
      wbc whole-body control

mindmap root((sonicstar)) What it does Text-to-robot-motion VLA inference Autonomous tasks Training pipeline Collect demonstrations Build dataset Train SonicLatent Deployment Policy server Simulation env Real robot future Repo structure starVLA training wbc whole-body control

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Train an AI policy that makes a Unitree G1 humanoid robot follow text instructions by learning from simulated human demonstrations.

USE CASE 2

Capture teleoperation data in simulation where a human operator controls the robot to build a training dataset for motion policies.

USE CASE 3

Deploy a trained Vision-Language-Action model to run tasks autonomously in simulation before attempting transfer to real hardware.

Tech stack

PythonNeural NetworkSimulationRobotics

Getting it running

Difficulty · hard Time to first run · 1day+

Requires NVIDIA GR00T-WholeBodyControl and starVLA as external dependencies, real-robot deployment on physical Unitree G1 hardware is a future milestone.

In plain English

SonicStar is an open-source project for training and deploying AI-controlled behavior on the Unitree G1 humanoid robot. The core idea is called Vision-Language-Action (VLA): you give the robot a plain-text instruction like "pick up the cylinder and throw it into the trash bin" and it executes the physical movements needed to follow that instruction, guided by camera input and a trained neural network. The repository is split into two main parts. The first (starVLA/) handles training and inference. You collect demonstration data from a simulation, train a model called SonicLatent on that data, then run a policy server that accepts text prompts and returns control signals. The second part (wbc/) handles the whole-body control side: running the simulation environment, capturing teleoperation data from a human operator, and connecting the trained policy to the robot's motors. The intended workflow is: run the simulation to gather training examples of a human completing tasks, process those recordings into a dataset, train the model, then deploy it back into the simulation (and eventually to the real robot) to run the task autonomously. The README includes keyboard shortcuts used during deployment to move the simulated robot through its initialization states before the policy takes over. The README is written entirely in Chinese and references two external codebases (NVIDIA's GR00T-WholeBodyControl and the starVLA framework) as the main systems this repository is built on top of. Real-world deployment on physical hardware is listed as a future milestone.

Copy-paste prompts

Prompt 1

I'm working with SonicStar for the Unitree G1 robot. How do I start the simulation to record teleoperation demonstrations, and what keyboard shortcuts control the robot during data capture in wbc/?

Prompt 2

I have collected demonstration data and want to train the SonicLatent model. What command do I run from the starVLA/ directory, and what format does the dataset need to be in?

Prompt 3

I have a trained SonicStar policy and want to test it in simulation before touching real hardware. How do I start the policy server and connect it to the wbc/ simulation environment so it accepts text prompts and sends control signals?

Prompt 4

SonicStar builds on NVIDIA's GR00T-WholeBodyControl and the starVLA framework. What do I need to install from each of those external codebases before I can run anything in this repository?

Open on GitHub → Explain another repo

← blackotters on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.