[Raspberry Pi logo] Raspberry Pi AI HAT+

Published October 2024

Overview

The Raspberry Pi AI HAT+ is an add-on board featuring a built-in [Hailo logo] AI accelerator designed for the Raspberry Pi 5. It offers an accessible, cost-effective, and power-efficient method for integrating high-performance Artificial Intelligence (AI) capabilities. This HAT enables exploration of applications such as process control, security systems, home automation, and robotics.

The AI HAT+ is available in two variants: 13 tera-operations per second (TOPS) and 26 TOPS. These variants utilize the Hailo-8L and Hailo-8 neural network inference accelerators, respectively. The 13 TOPS version is capable of running neural networks for tasks like object detection, semantic and instance segmentation, and pose estimation. The 26 TOPS version can handle more complex and larger neural networks, execute them faster, and manage multiple networks concurrently.

Connectivity is established via the Raspberry Pi 5's PCIe Gen 3 interface. When the host Raspberry Pi 5 is running an up-to-date Raspberry Pi OS, it automatically recognizes the on-board Hailo accelerator, making the Neural Processing Unit (NPU) available for AI computing tasks. The integrated rpicam-apps camera applications within Raspberry Pi OS natively support the AI module, automatically leveraging the NPU for compatible post-processing tasks.

Specification

Features:
  • Hailo-8 or Hailo-8L accelerator offering 26 TOPS or 13 TOPS inferencing performance, respectively.
  • Fully integrated into Raspberry Pi's camera software stack.
  • Conforms to the Raspberry Pi HAT+ specification.
  • Supplied with a 16mm stacking header, spacers, and screws for fitting onto a Raspberry Pi 5 equipped with the Raspberry Pi Active Cooler.
Operating temperature:
0°C to 50°C (ambient).
Production lifetime:
The Raspberry Pi AI HAT+ is planned for production until at least January 2030.
Compliance:
For a comprehensive list of local and regional product approvals, please visit pip.raspberrypi.com.

Physical Specification

The Raspberry Pi AI HAT+ has the following approximate dimensions in millimeters: Overall width of 65 mm, width between mounting holes of 58 mm, and length of 47.5 mm. A central area indicates the placement of the main AI accelerator chip, with space for a heatsink. A PCIe connector is visible at one end of the board. Side views illustrate the height of the board (approximately 17.5 mm) and the height of the mounting standoffs (around 16 mm). These standoffs are crucial for mounting the HAT onto a Raspberry Pi 5, especially when using the Raspberry Pi Active Cooler.

Note: All dimensions are approximate and for reference purposes only. They should not be used for production data. Dimensions are subject to manufacturing tolerances and may change.

Warnings

  • The Raspberry Pi AI HAT+ must only be connected to a Raspberry Pi via the PCIe interface and GPIO header.
  • Operate this product in a well-ventilated environment. If used inside a case, the case should not be covered.
  • While in use, this product must be firmly secured and should not be contacted by conductive items.
  • All peripherals used with this product must comply with relevant standards for the country of use and be marked accordingly to ensure safety and performance requirements are met.
  • Cables and connectors of all peripherals must have adequate insulation to meet relevant safety requirements.
  • Operation of this device requires adult supervision.

Safety Instructions

To prevent malfunction or damage to this product, please observe the following:

  • Do not expose to water or moisture, or place on a conductive surface while in operation.
  • Avoid exposing the product to heat from any source. Raspberry Pi computers and the Raspberry Pi AI HAT+ are designed for reliable operation at normal ambient temperatures.
  • Handle with care to prevent mechanical or electrical damage to the printed circuit board and connectors.
  • While powered, avoid handling the printed circuit board directly. Handle it by the corners only to minimize the risk of electrostatic discharge damage.

PDF preview unavailable. Download the PDF instead.

D0024060562 DocHub

Related Documents

Preview Raspberry Pi AI Kit: High-Performance AI Acceleration for Raspberry Pi 5
Discover the Raspberry Pi AI Kit, combining the Raspberry Pi M.2 HAT+ with a Hailo AI module for powerful, cost-effective AI acceleration on Raspberry Pi 5. Features a 13 TOPS NPU for applications in process control, robotics, and more.
Preview The MagPi Magazine Issue 143: Robot Explorers & Raspberry Pi Innovations
Explore cutting-edge robotics, AI, and electronics projects with The MagPi magazine, Issue 143. Featuring Robot Explorers, the Raspberry Pi AI Kit, coding tutorials, and inspiring success stories.
Preview Get Started with the Raspberry Pi AI Kit: A Beginner's Guide to AI and Edge Computing
Explore AI and edge computing with the Raspberry Pi AI Kit. This comprehensive guide by Dogan Ibrahim covers setup, software installation, Python programming, and practical AI projects, suitable for beginners.
Preview The Official Raspberry Pi Beginner's Guide - Contents
Explore the table of contents for The Official Raspberry Pi Beginner's Guide, covering essential topics from basic setup and programming in Scratch and Python to advanced hardware integration with Sense HAT and camera modules.
Preview Raspberry Pi AI Camera: Specifications and Overview
Detailed overview and technical specifications for the Raspberry Pi AI Camera module, featuring the Sony IMX500 Intelligent Vision Sensor for AI applications.
Preview The Official Raspberry Pi Handbook: Projects, Guides, and Reviews
Explore the world of Raspberry Pi with this comprehensive handbook. Discover projects, tutorials, and reviews covering the Raspberry Pi 5, Pico W, and various accessories. Learn electronics, build handheld devices, and delve into AI applications.
Preview Troubleshooting KMS HDMI Output on Raspberry Pi
A comprehensive guide to troubleshooting HDMI output issues with the KMS graphics driver on Raspberry Pi devices, covering common problems, symptoms, and mitigation strategies.
Preview Das offizielle Raspberry Pi Handbuch: Projekte, Anleitungen und Reviews
Entdecken Sie das Raspberry Pi 5 und Pico W mit diesem offiziellen Handbuch. Enthält detaillierte Projekte, praktische Anleitungen für MicroPython und KI, sowie Produkt-Reviews.