Raspberry Pi AI HAT+
The Raspberry Pi AI HAT+ is an add-on board designed for the Raspberry Pi 5 and includes an integrated Hailo AI accelerator.
When the AI HAT+ is attached to the Raspberry Pi 5, it enables rapid development of advanced AI vision applications that run in real time with low latency and low power consumption. Advanced neural networks such as object detection, semantic and instance segmentation, pose estimation, and facial feature recognition run solely on the Hailo coprocessor, freeing the Raspberry Pi 5 CPU for other tasks.
This add-on board provides an easy, cost-effective, and energy-efficient solution for integrating high-performance AI into a variety of applications, such as process control, security, home automation, and robotics.
Package contents
- AI HAT+ with built-in Hailo AI module containing the neural processing unit (NPU)
- Mounting hardware kit
- 16 mm stacking GPIO connector
Note: Raspberry Pi 5 is not included in the package (but can be purchased separately).
Features
- Hailo-8 or Hailo-8L accelerator, providing 26 TOPS or 13 TOPS inference performance respectively
- Fully integrated into the Raspberry Pi camera software stack
- Compliant with Raspberry Pi HAT+ specifications
- Includes 16 mm stacking connector, spacers, and screws for easy installation onto the Raspberry Pi 5 with the Raspberry Pi Active Cooler
About the AI HAT+
The AI HAT+ is available in two versions offering 13 and 26 tera-operations per second (TOPS), built around the Hailo-8L and Hailo-8 neural network inference accelerators. The 13 TOPS version efficiently runs neural networks for applications such as object detection, semantic and instance segmentation, pose estimation, and more. The 26 TOPS version supports larger networks, faster processing, and concurrent execution of multiple networks.
The AI HAT+ connects via the Raspberry Pi 5 PCIe Gen 3 interface. When the Raspberry Pi 5 is running an up-to-date Raspberry Pi OS image, it will automatically detect the Hailo accelerator and make the NPU available for AI compute tasks. The Raspberry Pi OS’s built-in rpicam applications natively support the AI module, using the NPU for compatible post-processing tasks.
Links