Gumstix announced the release of four Edge AI devices for machine-learning applications moving data from the network’s edge. The devices, which are powered by the NVIDIA Jetson Nano running a Quad-core ARM A57, feature built in TensorFlow support to make edge AI prototyping more accessible for software engineers.

According to the company, the AI devices streamline the software-hardware integration and provide users the following benefits:

- Easy to customize modular designs available in the Geppetto app.

- GPU enabled TensorFlow support.

- Customizable OS (Yocto) for rapid and simple deployment.

The Gumstix Snapshot have up to 16 video streams, along with the 3 other new boards.

Information from Gumstix about the boards are listed:

Gumstix Jetson Nano Snapshot Board

• 4 x Jetson Nano Vertical Connectors
• 16 x 15-pin Vertical Camera Connectors (Compatible with Raspberry Pi Camera Module V2)
• RJ45 Jack with Gigabit Magnetics
• Micro B USB Plug for flashing and another for serial debugging
• ESP32 for Jetson Nano, power and switch management
• Redundant WiFi backup connection for Ethernet failure
• Serial debugging available via USB connector or via network management

Gumstix Jetson Nano Development Board

• HDMI - USB 3.0 - 2 x Raspberry Pi Vertical Camera Connector
• M.2 (Key E) Wireless card interface
• Ethernet Connector
• USB Micro-B Jack for flashing
• USB Micro-B Jack for console debugging
• Raspberry Pi Compatible 40-pin header

Gumstix Jetson Nano MegaDrive

• HDMI - USB 3.0
• 2 x Raspberry Pi Vertical Camera Connector
• M.2 (Key M) SSD card interface
• 802.11 ac WiFi
• USB Micro-B Jack for flashing
• USB Micro-B Jack for console debugging

Gumstix Jetson Nano FastFlash

• Gumstix Jetson Nano FastFlash is a compact, cost-effective expansion board designed to quickly initialize or overwrite the Jetson Nano eMMC storage with a new disk image or mount the file system to your PC.

For more information, visit https://www.gumstix.com/