This week we’ll be giving away one Jetson Nano 2 GB Developer Kit from NVIDIA.

This version has 2 GB of memory, instead of the 4 GB on the original Nano, as well as one less USB port and two fewer MIPI CSI camera ports. But the treat here is that has allowed NVIDIA to drop the price of the Jetson Nano 2 GB to just $59. You hardly sacrifice anything in terms of performance. In fact, it’s even possible for training full AI models directly on the kit via transfer learning.

The Jetson Nano 2 GB Developer Kit is designed around a 128-core NVIDIA Maxwell GPU and quad-core Arm A57 CPUs, which connect to the 2 GB of 64-bit LPDDR4 memory at a blazing 25.6 GBps. This, of course, comes in super handy in demanding edge inferencing applications like computer vision that need to access memory frequently. To that end, so do the H.264 and H.265-compatible video encoder and decoder that support single-channel 4Kp30, four 1080p30 channels, or nine 720p30 channels and single channel 4Kp60, dual-channel 4Kp30, eight-channel 1080p30, or 18 channels at 720p30 resolution, respectively. 

This kit, which was developed for anyone new to artificial intelligence, robotics, and can be tuned to run in either 5W or 10W operating modes, features a wide range of I/O and connectivity, including:

• An 802.11ac wireless adaptor
• An HDMI port
• The aforementioned MIPI CSI-2 camera connector
• A Gigabit Ethernet port
• One USB 3.0 port
• Two USB 2.0 ports
• A Micro-USB port
• A USB Type C Connector that Supports a 5V power input
• A 40-pin expansion header for
• A 12-pin header that carries power and related signals
• A 4-pin header for connecting a fan in the event that the onboard heatsink isn’t enough to cool your application
• And a MicroSD card slot for storage

The microSD card usually isn’t included with the kit, but we’ll be shipping out one that includes a 64 GB microSD card with read speeds up to 100 MB/s and write speed up to 40 MB/s.

The Jetson Nano 2 GB is packaged with NVIDIA’s JetPack SDK that comes with an Ubuntu Linux distro, Linux Driver Package, support for the CUDA parallel programming model, and APIs and libraries for AI programming tools like TensorRT, cuDNN, OpenCV, and Vulkan 1.2. The kit also supports popular frameworks like Caffe2, Keras, MxNet, PyTorch, and TensorFlow. All of these are accessible via docker containers that can be downloaded from the Jetson Zoo. Advanced users can also take advantage of NVIDIA’s Python wheels.

If you’re interested in getting your hands on one of these, Jetson Nano 2 GBs just became available at the aforementioned $59 price point. But we have a couple here that we’ll be raffling off for free, per normal. All you have to do is register for this week’s raffle by filling out the form that’s on the other side of the link on your screen and in the description below.

Thanks for watching Dev Kit Weekly, and we’ll see you next week.