• Raspberry Pi AI HAT+2 kit box and Hailo-10H accelerator board for Raspberry Pi 5
  • Benchmark chart comparing local visual language model latency with and without AI HAT+2 accelerator
  • AI HAT+2 use case collage: local LLM/VLM demos, object detection, posture estimation, segmentation
  • Raspberry Pi 5 AI HAT+ series comparison table showing AI HAT vs AI HAT+ and AI HAT+ 2 specs
  • Tutorials and user guide overview for Hailo-10H AI HAT+2, with link and document thumbnails
  • Raspberry Pi AI HAT+2 with box on desk, highlighting 8GB onboard RAM and up to 40 TOPS text
  • Angled top view of Raspberry Pi AI HAT+2 board showing Hailo chip, ribbon connector, and GPIO header
  • Top-down view of AI HAT+2 PCB showing component layout and 40-pin GPIO stacking header
  • Back side of AI HAT+2 PCB showing certification marks, test pads, and serial label area
  • AI HAT+2 installed on Raspberry Pi 5 with heatsink and standoffs, showing low-profile fit
  • Package contents: AI HAT+2 board, heatsink, GPIO header, standoffs, screws, and quick start sheet
  • Raspberry Pi 5 AI HAT+2 Official Kit layout with Hailo-10H board and included mounting hardware
1 12

Hailo-10H AI HAT+2 Official Kit for Raspberry Pi 5, 40 TOPS (INT4) NPU, 8GB Memory, PCIe Gen3

Hailo-10H AI HAT+2 Official Kit for Raspberry Pi 5, 40 TOPS (INT4) NPU, 8GB Memory, PCIe Gen3

Yahboom

$414.00 USD
$414.00 USD
特價 售罄
已包含稅額。 結帳時計算運費。
  • PayPal
  • Visa
  • Mastercard
  • American Express
  • JCB
  • Discover
  • Diners Club
  • Google Pay
  • Apple Pay
查看完整資訊

Overview

The Hailo-10H AI HAT+2 Official Kit for Raspberry Pi 5 is a dedicated AI accelerator board designed for edge generative AI and computer vision workloads. It integrates a built-in Hailo-10H NPU rated at 40 TOPS (INT4) and includes 8GB of dedicated onboard memory, helping run large language models (LLMs) and vision-language models (VLMs) locally while leaving Raspberry Pi 5 system memory available for other tasks.

Key Features

  • Compatible with Raspberry Pi 5
  • Hailo-10H AI accelerator delivering 40 TOPS (INT4) inferencing performance
  • Onboard 8GB memory (dedicated)
  • Conforms to Raspberry Pi HAT+ specification
  • Connectivity via Raspberry Pi 5 PCI Express interface (PCIe Gen3 noted)
  • High compatibility with Raspberry Pi camera software stack: libcamera, rpicam-apps, Picamera2
  • Comprehensive software package / toolchain support (Hailo software components referenced: Hailo Model Zoo, Hailo Dataflow Compiler (SDK), HailoRT, TAPPAS, Hailo Firmware)
  • Operating temperature: 0C~50C (ambient)
  • Includes heatsink; supports use with an active cooler for improved ventilation (active cooler not included)

Specifications

Host Raspberry Pi 5
Accelerator chip (Hailo NPU) Hailo-10H
AI performance 40 TOPS (INT4)
Onboard memory 8GB
PCIe interface Raspberry Pi 5 PCIe Gen3 (standard compatible)
HAT interface Compliant with Raspberry Pi HAT+ specifications
Supported operating system Raspberry Pi OS
Supported frameworks (listed) TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch
Camera software stack support libcamera, rpicam-apps, Picamera2
Operating temperature 0C~50C (ambient)
Size 65 x 56.5 x 14 mm (with heatsink)

Performance References (from provided test figures)

  • Local VLM (Qwen2-VL-2B) response time (Raspberry Pi 5 4GB/8GB): without accelerator: >6 minutes (Pi 5-4GB), >3 minutes (Pi 5-8GB); with AI HAT+2: 3~30 seconds
  • YOLOv5 target detection: Raspberry Pi 5 test FPS: 8; Raspberry Pi 5 + AI HAT+2 test FPS: 30.01
  • Posture estimation: Raspberry Pi 5 test FPS: 1; Raspberry Pi 5 + AI HAT+2 test FPS: 30.64
  • Background segmentation: Raspberry Pi 5 test: unable to run (FPS: 0); Raspberry Pi 5 + AI HAT+2 test FPS: 29.63
  • Image detection: Raspberry Pi 5 test FPS: 1; Raspberry Pi 5 + AI HAT+2 test FPS: 60.22

What's Included

  • AI HAT+2 board (Hailo-10H)
  • Heatsink
  • 16mm stacking header
  • Support post(s)
  • Screws

Raspberry Pi 5 and an active cooler are not included. For integration or compatibility questions, contact support@rcdrone.top or visit https://rcdrone.top/.

Applications

  • Offline LLM and VLM deployment on Raspberry Pi 5
  • Real-time computer vision and image processing acceleration
  • Robotics
  • Offline process control and secure data analysis
  • Facilities management

Tutorials

  • Hailo10
  • Tutorial topics shown: Product Introduction; Local Deployment Tutorial for LLM and VLM; AI Vision Model Acceleration Tutorial; Raspberry Pi 5 AI HAT+2 Test Image
  • Local deployment topics shown: Environment Installation; Text-based interactive LLM; Visual and text interaction VLM
  • AI vision acceleration topics shown: Environment Setup; rpicam-apps; Test case; Instance segmentation; Pose estimation; Segmented application; Monocular depth
  • User Guide (listed)

Details

Raspberry Pi 5 Hailo-10H AI HAT+2 Official Kit board with 40 TOPS INT4 NPU and 8GB memoryKey feature icons for the AI HAT+2 kit: Raspberry Pi 5 support, 40 TOPS, Hailo-10H, 8GB memory, software toolsProduct introduction overview for the Raspberry Pi 5 AI HAT+2 with Hailo-10H accelerator and kit contents

With a dedicated Hailo-10H NPU and onboard 8GB memory, the kit helps run local LLM/VLM and vision workloads on Raspberry Pi 5.

Features list and top view of Hailo-10H AI HAT+2 board showing HAT+ connector and component layoutRaspberry Pi 5 AI HAT+ series comparison table highlighting AI HAT+ vs AI HAT+2 performance and memory

A quick comparison of the AI HAT+ lineup makes it easy to choose between the base model and the higher-performance AI HAT+2.

Local VLM benchmark chart comparing response time on Raspberry Pi 5 with and without the AI HAT+2 accelerator

Local visual-language model response times can drop from minutes to seconds when workloads are offloaded to the AI accelerator.

AI vision acceleration results for YOLOv5 detection and posture estimation, comparing Raspberry Pi 5 vs Pi 5 + AI HAT+2

Computer vision pipelines like object detection and pose estimation see large FPS gains when running with the AI HAT+2.

Background segmentation and image detection FPS comparison: Raspberry Pi 5 alone vs Raspberry Pi 5 with AI HAT+2Tutorial resource links and folders for local deployment guide covering LLM and VLM setup and test projects

Step-by-step tutorial materials are provided for setting up local LLM/VLM deployment and running sample projects.

AI vision model acceleration tutorial modules and user guide thumbnails for rpicam apps, segmentation, depth and moreHeatsink installation guide for AI HAT+2 showing thermal pad preparation before mounting to the accelerator board

Proper heatsink installation helps maintain stable performance during sustained AI inference workloads.

Heatsink positioning and fastening steps on the AI HAT+2 board using push pins and alignment pointsExploded installation view of AI HAT+2 on Raspberry Pi 5 with PCIe cable, pin header, heatsink and optional active cooler

For improved cooling under continuous load, the board supports pairing with an active cooler (Raspberry Pi 5 and cooler not included).

Hailo AI chip dataflow architecture diagram showing model compiler, parser and supported ML frameworksHailo software toolchain overview including Model Zoo, Dataflow Compiler SDK, HailoRT, TAPPAS and firmware

The software stack includes developer tools for model conversion, deployment, and runtime execution on the Hailo accelerator.

Industry application collage for edge AI, including voice assistant, translation, robotics, monitoring, and smart manufacturingLocal LLM and local VLM demo screenshots illustrating on-device text generation and vision-language interaction

Example demos include on-device text generation and vision-language workflows for offline edge AI prototyping.

Computer vision demo results for object detection, image segmentation, and posture estimation with console output and overlays

Typical vision tasks such as detection, segmentation, and pose estimation can be accelerated for real-time performance on Raspberry Pi 5.

Dimension drawing of the Hailo-10H AI HAT+2 board and radiator/heatsink, with measurements in mm.

The AI HAT+2 board and radiator dimensions in millimeters help confirm clearance and mounting fit in your Raspberry Pi 5 setup.

Shipping list for Raspberry Pi 5 AI HAT+2 kit with board, official heatsink, 16mm stacking header, and screws

The Raspberry Pi 5 AI HAT+2 kit includes the AI HAT+2 board, an official heatsink, a 16mm stacking header, and a screw package for installation.

Hailo AI HAT+2 kit board for Raspberry Pi 5 with 40-pin header, heatsink, ribbon cable and screws

The Hailo AI HAT+2 kit for Raspberry Pi 5 includes the add-on board, heatsink, ribbon cable, and mounting hardware for a clean install.

Related Collections