Skip to content

Releases: NVIDIA-AI-IOT/jetbot

Support JetPack 4.5

02 Feb 23:03
3f67669
Compare
Choose a tag to compare
Platform JetPack Version JetBot Version Download MD5 Checksum
Jetson Nano 2GB 4.5 0.4.3 jetbot-043_nano-2gb-jp45.zip e6dda4d13b1b1b31f648402b9b742152
Jetson Nano (4GB) 4.5 0.4.3 jetbot-043_nano-4gb-jp45.zip 760b1885646bfad8590633acca014289

Changes

Added

Added LocalController class which allows users to directly connect controller to JetBot

Fix for 3rd party motor variant

25 Jan 21:13
03f8ae6
Compare
Choose a tag to compare

Adds a small fix to enable later revision of waveshare JetBot.

Platform JetPack Version JetBot Version Download
Jetson Nano 2GB 4.4.1 0.4.2 jetbot-042_nano-2gb-jp441.zip
Jetson Nano (4GB) 4.4.1 0.4.2 jetbot-042_nano-4gb-jp441.zip

Jetson Nano 2G / Docker

25 Jan 20:25
Compare
Choose a tag to compare

This release is the first to support JetBot 2G as well as facilitate docker.

Platform JetPack Version JetBot Version Download
Jetson Nano 2GB 4.4.1 0.4.1 jetbot-041_nano-2gb-jp441.zip
Jetson Nano (4GB) 4.4.1 0.4.1 jetbot-041_nano-4gb-jp441.zip

Support JetPack 4.3

16 Jan 04:54
65d9eec
Compare
Choose a tag to compare

JetBot SD card image based on JetPack 4.3 (required for B01 revision of Jetson Nano Developer Kit ):

https://drive.google.com/open?id=1G5nw0o3Q6E08xZM99ZfzQAe7-qAXxzHN

Object Detection Engine:

https://drive.google.com/open?id=1KjlDMRD8uhgQmQK-nC2CZGHFTbq4qQQH

v0.3.0

10 Mar 07:25
Compare
Choose a tag to compare

SD Card image

Found here: jetbot_image_v0p3p0.zip

  • PyTorch 1.0 pre-installed
  • TensorFlow 1.13 pre-installed
  • 4GB swap enabled
  • default power model 5W
  • Jupyter Lab service installed
  • Status display service installed

Python package

  • Heartbeat class to determine disconnects from browser
  • Camera class to generate eventful camera frames
  • Robot class for controlling robot locomotion
  • Motor class for event based motor control
  • ObjectDetector class for detecting objects using the subpackage jetbot.ssd_tensorrt
  • TRTModel class for abstracting TensorRT model execution on numpy arrays

Notebooks

Robot

  • basic motion teaches programmatic robot control
  • teleoperation demonstrates video streaming and gamepad control
  • collision avoidance
    • data collection to collect free and blocked labels
    • live demo to demonstrate simple wander / avoid obstacles
  • object following
    • live demo to demonstrate multi network COCO object following and simultaneous collision avoidance

Host

  • collision avoidance
    • training to demonstrate training image classification network with PyTorch