BrainChip, the neuromorphic AI specialist, has announced on Twitter that it is now taking orders for two new development kits for its Akida advanced neural networking processor, as spotted by CNX Software. One of the development kits comes with a Raspberry Pi 4, and the other, which is twice the price, comes with a barebones X86 mini PC.
The Raspberry Pi in question is the Compute Module 4, with up to 8GB of RAM and up to 32GB flash-based storage (or an external Micro SD card if you choose the CM4 Lite), slotted into the official Raspberry Pi carrier board, which adds a PCIe slot and two HDMI outputs. The Akida ADK1000 AI accelerator then attaches itself to the PCIe slot, and the whole thing slots neatly into an enclosure.
If you’re feeling flush and fancy the PC version, you get something based on a Shuttle PC, with an LGA1200 socket and Intel 410 chipset ready for 10th-gen Intel processors and support for 32GB of RAM - you have to supply your own. There's an M.2 slot, and space for an additional SATA drive. The Akida card, once again, goes in a PCIe slot.
The ADK1000 can emulate 1.2 million neurons and 10 billion synapses in a spiking neural network. If that’s not enough for you, up to 64 of the devices can be daisy-chained together over PCIe. It’s designed for use as a stand-alone embedded accelerator or as a co-processor, and includes interfaces for ADAS sensors, audio sensors, and other IoT sensors.
A Brainchip press release, possibly written by an AI, states: “The solution is high-performance, small, ultra-low power and enables a wide array of edge capabilities. The Akida and intellectual property, can be used in applications including Smart Home, Smart Health, Smart City and Smart Transportation. These applications include but are not limited to home automation and remote controls, industrial IoT, robotics, security cameras, sensors, unmanned aircraft, autonomous vehicles, medical instruments, object detection, sound detection, odour and taste detection, gesture control and cybersecurity. Akida brings AI processing capability to edge devices for learning, enabling personalization of products without the need for retraining.”
If you’d like to know more, we suggest starting with the user guide.
