This is an amazing 5th of October 2020! A new Jetson Nano is released from NVIDIA!
One year and a half ago (March 2019) NVIDIA introduce the first NVIDIA Jetson Nano Developer kit 4Gb version, and now a new revolution is on the air!
The Jetson Nano Developer Kit became the go-to option for creating AI projects at the edge, and new developers began exploring more complex projects that were previously beyond their reach. Tens of thousands of developers and enthusiasts have adopted the Jetson Nano Developer kit and are actively contributing to the Jetson developer community with open-source projects, how-to’s, and videos.
The new NVIDIA Jetson Nano 2GB Developer Kit, priced at $59, makes it even more affordable for students, educators, and enthusiasts to learn AI and robotics
Pre-order Now: https://nvda.ws/30v5w3M
Jetson Nano 2GB Developer Kit is ideal for teaching, learning, and developing AI and Robotics applications for the following reasons:
- It is purpose-built for educators, learners, and enthusiasts
- Offers unbeatable price and performance for learning AI
- Delivers a learning journey from novice to expert
- Enables continual learning through easy to build Jetson Community projects
- Offers Jetson AI Workshop and Certification Programs
Many of the favorite sensors of the AI community such as the recently launched Raspberry Pi High Definition camera, Intel Real-sense camera, ZED 3D camera, other USB cameras, and WiFi dongles work out-of-the-box on Jetson Nano 2GB Developer Kit . This robust out-of-the-box support for
popular peripherals enables learners to focus on developing their projects rather than waste time on installation and debug of drivers.
Package and Unboxing
The new NVIDIA Jetson Nano 2Gb has the same packaging of the big Jetson Nano 4gb, the first difference that you notice is the new carrier board, with USB-C power plug compare the previous jack on the Jetson Nano 4Gb and other difference are listed below:
|NVIDIA Jetson Nano 2Gb Carrier||NVIDIA Jetson Nano 4Gb Carrier|
|USB-C power plug||Jack power plug|
|1 HDMI connector||1 HDMI connector|
1 DVI Connector
|3 USB3 ports||4 USB3 ports|
|1 Camera port||2 Camera ports|
|–||1 M.2 connector|
But there is great news!
The Jetson Nano 2Gb carrier is compatible with the NVIDIA Jetson Nano 4Gb carrier!
If you already have an NVIDIA Jetson 4Gb you can change the carrier and use on your Jetson Nano 2Gb. (PLEASE NOTE: you can NOT use the NVIDIA Jetson Xavier NX carrier on your Jetson Nano 2Gb)
Architecture and specifications
The NVIDIA Jetson Nano 2Gb use the same architecture of the 4Gb version, the main difference is the Memory size, only 2Gb. The NVIDIA Jetson Nano use a ARM57 CPU with a NVIDIA Maxwell 128core GPU, but below a short list of the main specifications for this developer kit:
|GPU||128-core NVIDIA Maxwell™|
|CPU||64-bit Quad-core ARM A57 (1.43 GHz)|
|Memory||2 GB 64-bit LPDDR4 (25.6 GB/s bandwidth)|
|Wireless connectivity||Available via an accessory 802.11ac wireless adaptor|
|USB||1x USB 3.0 Type A ports, 2x USB 2.0 Type A ports, 1x USB 2.0|
|40-Pin Header||GPIOs, I2C, I2S, SPI, PWM, UART|
|Camera||1x MIPI CSI-2 connector|
|Storage||MicroSD (Card not included)|
|Other IO||12-pin header (Power and related signals, UART)|
4-pin Fan header
|Size||100mm x 80mm x 29mm|
|Power||USB-C port (power brick not included)|
The NVIDIA Jetson Nano 2Gb such as the Nano 4gB has 2 NVP models:
Difference between Nano 4Gb & 2Gb
The NVIDIA Jetson Nano 2GB is an excellent entry-level Jetson board. I was surprised when I switched on in my first time, for the speed by LXDE desktop environment I used for few days like and home media centre, running video, browsing with chromium and writing with the text editor few notes.
Compare Unity, the speed and the performance of LXDE on a Jetson Nano 2GB are the same or better than an NVIDIA Jetson Nano 4GB.
I really appreciated being able to take advantage of the qualities of the LXDE graphics compared to Unity for the low memory consumption and the speed of response.
Compared to the NVIDIA Jetson Nano 4Gb with Unity, disable the desktop and work only with the command line, did not present big performance differences.
The only personal negative point, I did not use before the LXDE desktop and I spent a lot of time finding all simple application to work with it.
Another interesting note is the Code Name evolution, we are following the names of actors or plants in the last Star Wars trilogy. Yes is silly, but NVIDIA engineers for this family of board use characters in Star Wars! 🙂
|Jetson Nano 4Gb||Jetson Nano 2Gb|
When have seen in the Jetson Xavier NX the code name was Jakku the homeworld of the scavenger Rey.
Now the Jetson Nano 2Gb code name Batuu. It was a remote terrestrial planet on the edge of the galaxy‘s Outer Rim Territories, in the Batuu system of the Trilon sector. (more details in Wookiepedia)
The Jetson Nano 4Gb is the Porg were a species of sea-dwelling bird. They were native to the planet Ahch-To, where Jedi Master Luke Skywalker made his exile in the years prior to the Battle of Crait. More info from Wookiepedia
Jetson AI Certification Programs
The NVIDIA Deep Learning Institute (DLI) is launching the NVIDIA Jetson AI Certification Program, a series of hands-on self-paced educational tutorials,
video walkthroughs, and project-based assessments aimed at educators and learners. The Jetson AI Fundamentals course teaches the fundamentals of training/inference workflow, data collection, and real-time computer vision. It covers classification and regression networks, along with object detection
and semantic segmentation.
The course also covers building autonomous robots using JetBot and
implementing road following, collision avoidance, and object following:
There are currently two certifications included in the program:
- The Jetson AI Specialist certification for students, makers, hobbyists, or anyone looking to get started with AI.
- The Jetson AI Educator certification for educators and instructors who may want to teach AI in their own classes or courses.
Jetson Community Projects
There are amazing NVIDIA Jetson projects available, yes, sure, I kindly suggest first for all my jetson-stats to read the status of your board, but as well there are other friends that are made really cool projects, below I list some projects that I suggest:
BY RAFFAELLO BONGHI
Jetson-stats is a package to monitoring and controls your NVIDIA Jetson [Xavier NX, Nano, AGX Xavier, TX1, TX2] Works with all NVIDIA Jetson ecosystem.
When you install jetson-stats are included:
|FEVER CONTROL WITH JETSON NANO & LEPTON3|
BY WALTER “MYZHAR” LUCETTI
A useful application for the COVID19 era to control the human temperature and issue alarms in case of fever. This year, the year of COVID19, I decided to get that project out of the drawer and to adapt it to Nvidia Jetson Nano to realize an application to control human body temperature and issue alerts in case of fever.
|REAL-TIME HUMAN POSE ESTIMATION|
This project features multi-instance pose estimation
accelerated by NVIDIA TensorRT. It is ideal for applications where low latency is necessary. It includes:
– Training scripts to train on any key point task data in MSCOCO format
– A collection of models that may be easily optimized with TensorRT using torch2trt
This project can be used easily for the task of human pose estimation or extended for something new.
|RECOGNIZING SIGN LANGUAGE WITH JETSON NANO|
BY DENNIS FAUCHER
The Jetson Nano caches this model into memory and uses its 128 core GPU to recognize live images at up to 60fps. That high fps live recognition is what sets the Nano apart from other IoT devices. I have been hearing recommendations toward “Train in the cloud, deploy at the edge” and this seemed like a good reason to test that concept. Mission accomplished.
Running inference benchmarks is easy with our benchmarking utility hosted at https://github.com/NVIDIA-AI-IOT/jetson_benchmarks.git .
The utility provides a simple interface to run benchmarks.
Once the script starts, it will ask for a password two times, first for setting up the Jetson module in MAXN mode and second for setting jetson_clocks. The benchmarking will take a couple of hours and will run benchmarks on all the models. At the end, you will see benchmarking results similar to this:
Jetson Nano 2GB Developer Kit supports a diverse set of AI models and frameworks and delivers best-in-class performance demanded by AI practitioners and engineers today. Other alternatives in the market like Raspberry Pi-4 or Google Coral just do not have the performance that current AI projects demand. NVIDIA Jetson Nano 2GB Developer Kit offers up to 200X the performance of Raspberry Pi 4 though being in a similar price range.
Below is a table comparing Jetson Nano 2GB Developer Kit features with Raspberry Pi 4 and Google Coral Development board.
|Features||Jetson Nano Developer Kit||Raspberry PI 4||Coral (Beta) Developer Board|
|CPU||Quad-core ARM A57||Quad-core ARM A72||Quad-core ARM A53|
|GPU||472 GFLOPS (FP16)||–||64 GFLOPS (FP16)|
|Memory||4GB 64 bit LPDDR4 1600MHz|
|2GB, 4GB or 8GB 64 bit|
LPDDR4 3200MHz SDRAM
(depending on model)
|1GB 32 bit LPDDR4|
|DL HW||CUDA GPU||CPU||TPU|
|DL SW||Supports all the popular frameworks|
Supports wide variety of models
Supports wide variety of model
|Supports only a simplified framework|
– TensorFLow Lite
Supports only few selected models
|Codec HW||4K60 | 8x 1080p30 HEVC|
4K30 | 4x 1080p30 HEVC
|4K60 HEVC Decode|
1080P60 AVC Decode
1080P30 AVC Encode
|4K60 HEVC Decode|
No Encode Support
|WiFi/BT USB3 Ports||External WiFi Dongle bundled|
1 USB 3.0
2x USB 2.0
1x USB 2.0 Micro-B
2 USB 3.0
2 USB 2.0
1 USB 3.0
|Price||$59||Starting from $35||$129|
The below graph shows AI inference performance of Jetson Nano 2GB Developer Kit compared to Raspberry-Pi 4 and Google Coral.
Copyright © 2020. All Rights Reserved.