Edge Impulse Partners with Microchip to Offer Support for On-Device AI/ML on the SAMA7G54
Efficiency-focused 32-bit Arm Cortex-A7 part now gets full Edge Impulse support, including compatibility with the FOMO algorithm.
On-device machine learning specialist Edge Impulse has announced a partnership with Microchip to bring the company's high-efficiency SAMA7G54 microprocessor into the fold — offering full support for building, training, and deploying models using the SAMA7G54 as a target.
"The Edge Impulse platform is designed to provide engineers with an easy-to-use device to develop AI/ML [Artificial Intelligence/Machine Learning] models," says Edge Impulse co-founder and chief executive officer Zach Shelby. "Our collaboration with Microchip to integrate the SAMA7G54 MPU into the platform aligns with our vision to empower developers to bring more AI products to market in weeks instead of months or years."
"This innovative solution leverages Microchip’s high-performance single-core MPU [Microprocessor Unit] and the Edge Impulse platform to enable developers to train, evaluate and deploy machine learning models," adds Microchip's Rod Drake. "The adoption of AI at the edge is rapidly increasing across many applications and this solution is designed to significantly streamline the process from the design phase through implementation."
The Microchip SAMA7G54 is a single-core 32-bit Arm Cortex-A7 microprocessor running at up to 1GHz and built with efficiency in mind, with Microchip focusing particularly keenly on the automotive market. The part incudes a built-in imaging and audio system supporting two-lane MIPI Camera Serial Interface (CSI) and 12-bit parallel camera connectivity for sensors up to eight megapixels at 60 frames per second, gigabit and Fast Ethernet MACs, High-Speed USB 2.0, CAN-FD, SD/MMC, quad- and octal-SPI, and up to 12 FLEXCOMs for peripheral connectivity.
The chip's integration into the Edge Impulse platform means that it's now possible to target the SAMA7G54 when building, training, and deploying models for on-device use — including Edge Impulse's Faster Objects More Objects (FOMO) algorithm, which is designed for peak performance on resource-constrained devices.
Information on getting started with Edge Impulse for the SAMA7G54 is available on the company's documentation site; information on Microchip's SAMA7G54-EK evaluation board is available on the official product page.