Note: It is highly recommended to get a proper cooling mechanism (Active or Passive) and a proper PD (Power Delivery 5A@5V) PSU for the Raspberry Pi 5 and the AI HAT+.
Official AI HAT+ for Raspberry Pi 5
To continue the successful partnership with Hailo, Raspberry Pi introduces the new Raspberry Pi AI HAT+! The Raspberry Pi AI HAT+ is an add-on board featuring a built-in Hailo AI accelerator for the Raspberry Pi 5. It offers an accessible, cost-effective, and power-efficient solution for integrating high-performance AI into your projects. Applications include process control, security, home automation, and robotics. This official product enables you to unlock the incredible potential of AI, especially when paired with the Raspberry Pi 5.
Unparalleled Performance with HAILO AI Accelerators
The AI HAT+ is available in 13 and 26 tera-operations per second (TOPS) variants, built around the Hailo-8L and Hailo-8 neural network inference accelerators. The 13 TOPS variant capably runs neural networks for applications including object detection, semantic and instance segmentation, pose estimation, and more. The 26 TOPS variant can handle larger networks, can run them more quickly, and can more effectively run multiple networks simultaneously.
Hailo accelerators go beyond raw power. It claims to be the most efficient edge AI processor in the market, achieving an order of magnitude better performance in both area and power consumption than competitors. This efficiency is crucial for running deep learning models on resource-constrained, battery-powered, edge devices like Raspberry Pi 5. Additionally, the Hailo accelerators utilize a specialized architecture optimized for neural networks which is the Neural Processing Unit (NPU). It operates within a wide temperature range and offers scalability to handle multiple data streams and models simultaneously, making it a versatile solution for various AI applications.
Here is the comparison that we (Cytron Technologies) list out, comparing the Hailo 8L - 13 TOPS AI Accelerator with other popular AI accelerators:
Seamless Integration and Hassle-Free Development
The AI HAT+ communicates using Raspberry Pi 5’s PCIe Gen 3 interface. When the host Raspberry Pi 5 is running an up-to-date Raspberry Pi OS image, it automatically detects the on-board Hailo accelerator and makes the NPU available for AI computing tasks. It also integrates seamlessly with Raspberry Pi OS and the existing camera software stack (libcamera, rpicam-apps, and picamera2). This means you can leverage existing libraries and tools within your familiar development environment, saving valuable time and streamlining the creation process. And of course, it is also quite straightforward to install the driver and software package (HailoRT, Hailo Tappas, and more) for the Hailo accelerators from the existing Raspberry Pi OS using the apt package manager.
Both versions of the AI HAT+ are fully backward compatible with the AI Kit. Existing Hailo accelerator integration in the camera software stack works in exactly the same way with the AI HAT+. Any neural network model compiled for the Hailo-8L will run smoothly on the Hailo-8; while models specifically built for the Hailo-8 may not work on the Hailo-8L, alternative versions with lower performance are generally available, ensuring flexibility across different use cases.
Check out some of the demos from Raspberry Pi (all are using Raspberry Pi 5 with the AI HAT+):
The HAILO AI Accelerator offers:
- Performance: Up to 26 TOPS (Tera Operations Per Second) processing capability, making it efficient for real-time AI tasks.
- Efficiency: Hailo claims to outperform all other edge AI processors in area and power efficiency by an order of magnitude.
- Architecture: Utilizes a structure-driven dataflow architecture to optimize neural networks.
- Power Consumption: Low power consumption allows for running deep neural network inference on edge devices.
- Temperature Range: Supports a wide temperature range of 0°C to 50°C for various environments.
- Scalability: Enables simultaneous processing of multiple data streams and models, making it suitable for complex AI applications.
Perfect for Computer Vision and Edge AI Robotic
BTW, the AI HAT+ is fully compatible with the latest Raspberry Pi OS (Bookworm) and also all the camera modules and the software stack. We are very excited about this innovation from Raspberry Pi, it empowers many of us to:
- Develop cutting-edge AI applications:
- Image recognition and classification
- Object detection and tracking
- Facial recognition
- Natural Language Processing (NLP) tasks (with additional software)
- And much more!
- Elevate your projects:
- Add intelligent features to your Raspberry Pi creations
- Experiment with real-world AI applications
- Push the boundaries of creativity and innovation
You will see a leap improvement in existing computer vision projects based on Raspberry Pi, for example, the Donkeycar, and OpenCV image processing projects. The Raspberry Pi AI HAT+ for Raspberry Pi 5 is the perfect solution for developers, hobbyists, and educators seeking to explore the exciting world of AI. Join the revolution and take your Raspberry Pi projects to the next level!
FAQ for Raspberry Pi AI HAT+
- Can I use the Raspberry Pi AI HAT+ with the Raspberry Pi 4 Model B?
Answer: No, because the AI HAT+ required a PCIe socket for the interface/connection and so far, only Raspberry Pi 5 has that socket. - Can I use Raspberry Pi OS (Buster or Bullseye) older versions of Raspberry Pi OS work with Raspberry Pi AI HAT+?
Answer: You will need the latest version of Raspberry Pi OS, Bookworm (64-bit) because the AI HAT is meant for Raspberry Pi 5, and Raspberry Pi 5 can only boot up with Bookworm. - Will my Raspberry Pi 4 power supply (3.0A or 3.5A @ 5V) work with the Raspberry Pi AI HAT+?
Answer: The official Raspberry Pi 15W USB-C PSU or other compatible PSU for Raspberry Pi 4 can technically power the Raspberry Pi 5 and the AI HAT+, it might not be sufficient due to the Raspberry Pi 5's increased performance, and not to forget the AI kit requirement. For optimal performance and stability, we strongly recommend the official 27W high-quality 5V 5A USB-C power supply for Raspberry Pi 5 and this AI HAT+. - Is the active cooler needed and fit with the AI HAT+?
Answer: Raspberry Pi 5 is faster and more powerful than prior-generation Raspberry Pis, and like most general-purpose computers, it will perform best with active cooling. The Raspberry Pi Heatsink & Active Cooler is a good cooling solution for the AI HAT+. - Can I boot up Raspberry Pi 5 from NVMe SSD while using the AI HAT+?
Answer: If you plan to use the PCIe bus for the NVMe SSD, then it will be NOT Possible because the AI HAT+ needs the PCIe bus. Yet, you can still get a USB 3.0 to NVMe M.2 SSD adapter and boot the Raspberry Pi 5 from the USB 3.0 port. - Will the Raspberry Pi AI HAT+ fit the Official Raspberry Pi 5 case?
Answer: Yes, but you will need to remove the top lid and the fan.
Warranty Period: 12 months
International
Singapore
Malaysia
Thailand
Vietnam
