AI Hat with Hailo 8 for Raspberry Pi 5 - PCIe

AI Kit for Pi 5 based on Hailo 8 module - 26 TOPS power

This artificial intelligence HAT (AI Kit) for Raspberry Pi 5 is based on a Hailo 8 neural accelerator, a component directly soldered onto the HAT.
This HAT is therefore equipped with the unbridled Hailo 8 version (and not the Light version as on the AI ​​kit which develops 13 Tops)
This Hat develops 26 TOPS, twice the power of the AI kit (13 Tops) with 26 TOPS, or 26 Tera operations per second available for inference calculations.

Thanks to the Hailo 8 AI component, this configuration offers an accessible artificial intelligence solution with a good efficiency-power ratio for Raspberry-Pi 5.
This HAT allows to explore advanced applications such as: human + posture identification, object identification, process control, smart security, home automation and robotics.

With that power level, the AI ​​accelerator has a processing power of 26 TOPS (26 tera-operations per second) for the inference of its neural network. Thanks to the PCIe communication channel of the Pi 5, the accelerator can efficiently transfer information with the cores of the Raspberry-Pi. This also optimizes the transmission of data from the camera modules.

With 26 TOPS, the following can be executed:

• the same Hailo 8L inference models but much faster (close to real time).
• more complex Hailo 8 models.
• multiple models at the same time (on the same data source).

It seems obvious that most applications will use the AI ​​module with the camera module (or two cameras for a stereoscopic view).
Many pre-trained AI models are available, this allows you to quickly take advantage of the HAT's capabilities. Note that the Hailo 8 module does not have enough power to allow the creation and training of your own models.

A version of Raspberry-Pi OS for Pi 5 will automatically detect the Hailo and the NPU (Neuronal Processing Unit) will be available for AI processing.
The rpicam-apps camera applications integrated into Raspberry Pi OS support the AI ​​module (native support), and will automatically use the NPU to perform post-processing tasks (tasks compatible with the NPU).

The AI ​​HAT provides a particularly efficient scheduler for the hardware accelerator. It allows to run multiple neural networks on a single camera -OR- one/several neural networks on two cameras (concurrent processing) see source.

Note: this product requires a Raspberry-Pi 5 and a camera to work.
Need to use GPIO with the AI ​​Kit? See our GPIO Replicator which can be soldered on the stacking header before placing the AI ​​Kit HAT.

What can we do with the kit ?

What can be done with the NPU depends on the available training models.
Fortunately, Hailo does provides lots of pre-trained public models on its hailo_model_zoo repository and Raspberry-Pi's AI Kit documentation will guide you throught the required steps.

The full list for Hailo 8L is available in this sub-page of the repository, some of them are documented in the Hailo Examples for Raspberry-Pi 5.

Here some demonstration video:

• Object Recognition in Traffic (Youtube)
• Object Recognition in Office (Youtube)
• Pose/posture Estimation (Youtube)
• Subject Segmentation (Youtube)

Jeff Geerling also published a very nice Pi AI video about the Hailo 8L kit for Raspberry-Pi 5.

Technical details

• For Raspberry Pi 5
• Hailo 8 accelerator with up to 26 TOPS inferencing performance
• HAT+ format
• Pre-assembled and tested (ready out of the box)
• Fully integrated into Raspberry Pi’s camera software stack
• Operating temperature: 0°C to 50°C (ambient)
• Can be fitted on Raspberry Pi 5 + Active Cooler

Content

For each order, you will receive the Hat with the AI
The bundle is supplied with a 16mm stacking header and spacers/screws. PCIe Ribbon cable included. See the assembling instruction in the AI Kit documentation.

The HAT can be fitted to a Raspberry Pi 5 with the Raspberry Pi Active Cooler in place.

Please note: Raspberry-Pi and Active Cooler are not included.