README.md

Jetson Nano

Tutorials

• Two days to a demo https://developer.nvidia.com/embedded/twodaystoademo

Utils

Get info about your board:

$ cat /proc/device-tree/model
NVIDIA Jetson Nano Developer Kit

Get information about the NVIDIA Jetson OS environment, L4T, JetPack version, using jetsonhacks/jetsonUtilities:

$ ./jetsonInfo.py 
 NVIDIA Jetson TX1
 L4T 32.2.0 [ JetPack 4.2.1 ]
 Ubuntu 18.04.4 LTS
 Kernel Version: 4.9.140-tegra
 CUDA 10.0.326

Frameworks

• Fast.AI

• Scripts for building OpenCV on the NVIDIA Jetson Nano

  • seems to need to install qt4-make or qt5-make (not sure which one)

• PyTorch & Torchvision

• Instructions to install TensorFlow on Jetson Nano

• OpenCV & dlib

• This guy also have his custom scripts to build opencv, ssd caffe, tensorflow & https://jkjung-avt.github.io/opencv-on-nano/

• Tensorflow

  • Installing TensorFlow For Jetson Platform (also as PDF)
  • Release Notes For Jetson Platform: See TensorFlow compatibility with NVIDIA containers and Jetpack
  • Prebuilt Tensorflow python wheels per Jetpack version

Camera

• Jetson Nano + Raspberry Pi Camera
• NVIDIA-AI-IOT/jetcam
  • See Configure Nvidia Jetson Nano as an AI testbed

Fan

• I've got the Noctua NF-A4x10 5V PWM 40x10mm Premium Fan
  • You'll need additional M3-16 (3mm,16mm) screws to mount it as the silicon screws provided by Noctua are too big for the Nano dissipator's holes (3mm).
• NVidia recommends the Noctua NF-A4x20 5V PWM 40x20mm Premium Fan

How to control the fan:

• Manually (0 off, 255 full speed):

sudo sh -c 'echo 255 > /sys/devices/pwm-fan/target_pwm'

• Or Automatically: Jetson Fan Control

Power Management

• Official Power Management for Jetson Nano page
• Max perf: sudo /usr/bin/jetson_clocks
• Power Mode (Mode 0 is 10W MAXN, Mode 1 is 5W): sudo nvpmodel -m 0

Case

• NanoBox Connectors' Edition

Things I've got to check

• Docker containers, Computer Vision and Cuda

Reference

• Nvidia Doc Jetson Modules and Developer Kits