At the ongoing GPU Tech Conference 2019, Nvidia has announced a new member in its Jetson family – Jetson Nano. It is a $99 AI computer that can power AI-powered robots and other devices.
Jetson Nano is a single-board computer in the form of a module. By making it available in a smaller form factor, Nvidia has tried to entice DIY enthusiasts who can use this module in their different creations.
Jetson Nano: Specifications
Jetson Nano is available in two form factors. A Nano developer kit at $99 aimed at DIY project makers and developers and a production-ready module for commercial organizations starting at $129.
The $99 AI computing dev kit offered by Nvidia features a 128 CUDA core GPU, based on the Maxwell architecture along with 472 GFLOPs. The device is powered by a 64-bit quad-core ARM Cortex-A57 based CPU and 4GB of RAM.
Nvidia Jetson Nano also carries a video processor that supports 4K 30fps encoding and 4K 60fps decoding. Additionally, it supports PCle components and has a USB 3.0 port as well.
The $129 Jetson Nano Module, which is meant for organizations, is a completely built module with power management, memory, fully accessible input/output ports and clocking.
It is aimed at reducing overall development time as companies can update performance and capabilities even after the system has been deployed.
JetBot: A DIY Robot Based On Jetson Nano
Nvidia has also announced JetBot — a Jetson Nano-based open-source AI project that will give users an idea of the capabilities of the single-board computer.
It comes with a robot chassis, motors and batteries for allowing users, familiar with Python, to build a DIY a self-driving robot.
Nvidia’s Jetson Nano also supports commonly used AI frameworks like Caffe, TensorFlow, Keras, PyTorch, and MXNet.
Do you think Nvidia has launched a Raspberry Pi killer in the form of Jetson Nano? Would you buy it? Tell us in the comments below.
Also Read: Intel & Facebook Are Collaborating To Design AI Deep Learning Processor