Start training an ACT or gr00t model on the specified dataset. This will upload a trained model to the Hugging Face Hub using the main branch of the specified dataset.

Start training a model

Training paramters are left to None by default and are set depending on the dataset in the training pipeline.

Batch size for training, we run this on an A10G, leave it to None to auto-detect based on your dataset

Batch Size

Number of training steps, leave it to None to auto-detect based on your dataset

Steps

TrainingParamsAct

Batch size for training, default is 64, decrease it if you get an out of memory error

Data Dir

Number of epochs to train for, default is 10

Epochs

Learning rate for training, default is 0.0001

Learning Rate

Output Dir

The path to the Isaac-GR00T repo. If not provided, will assume we are in the repo.

Path To Gr00T Repo

Train test split ratio, default is 1.0 (no split), should be between 0 and 1

Train Test Split

Optional directory to save the validation dataset to. If None, no validation will be done.

Validation Data Dir

Optional dataset repository ID on Hugging Face to use for validation

Validation Dataset Name

TrainingParamsGr00T

Location

Message

Error Type

ValidationError

Pydantic model for training request validation

TrainingRequest

Default response. May contain other fields.

StatusResponse

HTTPValidationError

phospho

An absurdly simple way to train AI models for real-world robots, built for ML engineers.

Welcome to phosphobot

How to set up phosphobot for free and control your robot arm in simulation and in real life.

Install phosphobot

How to remote control a robot arm with a keyboard, in VR, with an API or a leader arm?

Control your robot arm

How to record a robotics dataset with your robot?

Record a robotics dataset

How to repair, merge, split and delete LeRobot datasets

Manipulate datasets

How to train a robotics AI model with a dataset?

Train a robotics AI model

How to make an AI model control your robot?

Control robot with AI models

How to control your robot arm using a Meta Quest 3 over a local WiFi network

Teleoperation on local network

Core Robotics Concepts

How to move a robot arm to a specific position.

Kinematics

Learn about the LeRobot dataset format: its structure, versions, how to use it, and common tips.

LeRobot Dataset Format

Learn how to record a good robotics dataset.

Datasets Best Practices

A guide to training AI models that control robots

Train robotics AI models

How to Train and Run SmolVLA with LeRobot: A Step-by-Step Guide

Train SmolVLA

Policies in AI Robotics

Understanding Gravity Compensation in Robotics

Gravity Compensation

What are the different cameras used for robotics ?

Cameras

Unbox and set up your phospho starter pack.

Starter pack unboxing (DK2)

How to set up phosphobot and control your SO-100 robot arm

Quickstart with your SO-100

Dev kit unboxing (DK1)

How to build the SO-100 robot arm? Step by step guide

Build the SO-100 robot arm

What kind of telemetry does phospho collect?

Telemetry

How to control the robot arm with API calls.

Simple controls

How to leverage vision models for robotics.

Computer vision

Control your robot over the internet using ngrok

Teleoperation from anywhere in the world

Github

Get hardware

Contact us

Initialize the robot to its initial position before starting the teleoperation.

Initialize Robot

Teleoperation Control

High frequency control of the robot arm via WebSocket connection for real-time updates and commands.

Teleoperation Control (WebSocket)

Move the robot to an absolute position specified by the end-effector (in centimeters and degrees). Make sure to call `/move/init` before using this endpoint.

Move to Absolute Position

Move the robot to a relative position based on received delta values (in centimeters and degrees).

Move to Relative Position

Use the leader arm to control the follower arm.

Use the leader arm to control the follower arm

Stop the leader-follower control

Retrieve the position, orientation, and open status of the robot's end effector. Only available for manipulators.

Read End-Effector Position

Read the current torque of the robot's joints.

Read Torque

Enable or disable the torque of the robot.

Toggle Torque

Read the current positions of the robot's joints in radians and motor units.

Read Joint Positions

Move the robot's joints to the specified angles.

Write Joint Positions

Start the calibration sequence for the robot.

Calibrate Robot

Asynchronously start recording an episode in the background.
Output format is chosen when stopping the recording.

Start Recording Episode

Stop the recording of the episode. The data is saved to disk to the user home directory, in the `phosphobot` folder.

Stop Recording Episode

Play Recording

Start the auto control by AI

Stop the auto control by AI

Stream video feed of the specified camera. If no camera id is provided, the default camera is used. If the camera id is 'realsense' or 'depth', the realsense camera is used.Specify a target size and quality using query parameters.

Getting Started

phosphobot Basic Usage

Learn about AI and robotics

Hardware

API Reference

FAQ

Examples

Start training a model

Body

Response