device
-
Introducing edge AI solutions using Hailo-8 and the i.mx8QuadXPlus evaluation board.
Hailo-8/8L (AI Accelerator)
AI accelerator coprocessors that integrate with edge platforms are low-power, compact, and low-cost, enhancing real-time deep learning inference tasks at the edge.
Because it can be added externally to existing SoCs, development time can be reduced compared to changing to an AI-integrated SoC.
- Operates as a standalone device (no external DYNAMIC RAMS required).
- NPU maximum 26TOPS (8L: 13TOPS)
- Scalability (312 TOPS with 12 devices)
- TRANSFORMERS compatible
- Connects to any SoC/PC via PCIe.
- Low power consumption (Typ2.5W)
- AEC-Q100 grade 2 (-40℃~105℃)
Hailo-10H (AI Accelerator)
The Hailo-10H AI processor delivers a computing power of up to 40 teraoperations per second (TOPS), significantly outperforming other edge AI processors.
Its coverage area and power efficiency are orders of magnitude better than other major solutions.
Hailo-10H leverages a unique, powerful, and scalable structure-driven dataflow architecture to maximize the fundamental characteristics of neural networks.
This makes it possible to run deep learning applications on edge devices at full scale more efficiently, effectively, and at a significantly lower cost compared to conventional solutions.
- NPU up to 40TOPS
- TRANSFORMERS compatible
- Achieves high-performance processing of generative AI models, including LLMs, with minimal CPU/GPU load.
- Connects to any SoC/PC via PCIe.
- Low power consumption (Typ2.5W)
- AEC-Q100 grade 2 (-40℃~105℃)
Hailo-15 SoC (AI Vision Processor)
Hailo-15 is an AI vision processor for smart cameras.
The Hailo-15SoC (System-on-a-Chip) combines Hailo's patented and market-proven advanced AI inference performance with a sophisticated computer vision engine to deliver data-reach video analytics.
Unprecedented AI capabilities enable the execution of advanced AI-powered image processing and multiple complex deep learning AI applications at full scale and with extremely high efficiency.
- AP Quad-core ARM53@1.3GHz
- NPU max 20TOPS
- DSP Vector Processor 350GOPS
- AEC-Q100 grade 2 (-40℃~105℃)
Platform
Hailo-8 M.2 AI Acceleration Module
The Hailo-8 M.2 module is an AI accelerator module for AI applications and is compatible with the NGFF M.2 form factor (M, B+M, and A+E keys).
The AI module is equipped with a highly power-efficient Hailo-8 AI processor with 26 tera operations per second (TOPS).
The M.2 AI accelerator features a full PCIe Gen-3.0 2-lane interface (4 lanes in the M-key module), delivering unprecedented AI performance for edge devices.
The M.2 module can connect to existing edge devices equipped with an M.2 SOCKETS, enabling low-power, real-time deep neural network inference for a wide range of market segments.
By leveraging Hailo Technologies' cutting-edge dataflow compiler and support for standard AI FRAMES, customers can easily port neural network models to Hailo-8 and rapidly bring high-performance AI products to market.
Hailo-8 Century High-Performance PCIe Card
Today, intelligent vision systems require high resolution, high FRAMES rates, and multiple sensors, as well as real-time, high-performance, and accurate AI processing capabilities.
Hailo Technologies provides up to 208 TOPS of AI performance in a low-power PCIe form factor accelerator card to support developers of high-performance edge video analytics systems across multiple industries.
The high-performance Century PCIe card features 16 PCIe slots, enabling low-power, real-time deep neural network inference for a wide range of markets and applications on any platform.
Hailo-10H M.2 Generative AI Acceleration Module
Featuring an M.2 form factor, the Hailo-10H GenAI accelerator-based module can be easily inserted into existing PC and edge device M.2 SOCKETS, enabling real-time deep neural network inference across various applications and market segments with low power consumption.
Available Software Components
A wide range of software development environments
By registering on the Hailo Technologies website, you can download the following software.
You can evaluate the Hailo-8 M.2 Module immediately after purchasing it.
- Hloco Model Zoo
- Hailo Dataflow Compiler (SDK)
- TAPPAS App Examples & Pipelines
- HartoRT
- Hailo-8™ Firmware
application
Explore AI tasks on edge devices across a variety of applications and industries.
Advanced Driver-Assistance Systems (ADAS) and Autonomous Driving (AD) / Security / Industrial Automation / Retail / Personal Computing
Advanced driver-assistance systems (ADAS) and autonomous driving (AD)
AI-powered automated driving
Since the invention of the first automobile, safety has been a driving force behind innovation and progress in the automotive industry.
Recently, machine learning is being used not only to further enhance the safety of automobiles, but also to improve driver efficiency and convenience.
Machine learning is applied to both advanced driver-assistance systems (ADAS) and autonomous driving (AD). AI algorithms analyze and interpret data from cameras, radar, LiDAR, and other perceptual sources to navigate the road and make driving decisions.
These enhancements will be utilized in lightweight vehicles, motorcycles, bicycles, and even off-road vehicles, creating a variety of applications beyond highway driving.
Automated driving at the edge
From lane keeping to traffic congestion assistance, and ultimately to piloting systems for urban areas, we need AI-centric machines with advanced computing power.
The system needs to process input from multiple sensors to create reliable 360-degree situational awareness and aerial road views in real time. Furthermore, the system must meet several constraints, including heat dissipation, mass-production quality, size, and cost.
Hailo Technologies collaborates with leading OEMs and Tier-1 suppliers at the forefront of AI automotive technology, bringing cutting-edge deep learning performance to a variety of assistive and autonomous driving systems while complying with evolving safety standards and regulations.
The Hailo AI processor is an AEC-Q100 Grade 2 certified automotive AI accelerator developed in accordance with ISO26262 guidelines to meet ASIL-B compliance.
Security
Advanced Road Traffic Systems
Artificial intelligence (AI) enhances intelligent transportation systems (ITS) to streamline traffic, alleviate congestion, improve safety, and warn of traffic violations and dangerous situations.
By applying AI models, it becomes possible to process images and videos of traffic conditions in real time with high accuracy, and to detect and track objects such as vehicles and pedestrians.
Boundary protection
Boundary protection refers to security measures and systems implemented to protect the physical boundaries of a place or property.
The primary purpose of perimeter protection is to prevent unauthorized Access, detect potential threats, and stop intruders from entering the protected area.
Industrial automation
Automated Optical Inspection (AOI)
Automated optical inspection (AOI) is a cutting-edge technology that is revolutionizing industrial automation processes.
AOI (Automated Optical Inspection) is used in various manufacturing sectors and plays a crucial role in ensuring product quality and improving production efficiency and productivity. By utilizing high-resolution cameras and advanced image processing algorithms, AOI systems quickly and accurately inspect and identify all manufactured components.
Optical inspection can be used to scan and identify manufactured goods, classify components, or identify defects such as surface anomalies, misalignments, or missing features. This real-time, non-destructive testing enables rapid identification and correction of defects, minimizing costly rework and product recalls.
Furthermore, it can handle large volumes of parts with unparalleled precision, streamlining production lines and enabling industries to achieve higher productivity, superior quality, and increased customer satisfaction. As technology continues to evolve, automated optical inspection will play an even more crucial role in shaping the future of industrial manufacturing.
retail
Smart Retail
Smart retail is reshaping the traditional shopping environment by leveraging AI to not only improve the shopping experience but also to help businesses maintain their competitiveness and appeal, in light of increasing competition from online shopping platforms. AI is used for the following purposes:
- Product and individual identification streamlines the shopping process, enabling fast, accurate, and efficient checkout.
- By analyzing vast amounts of customer data, retailers can understand consumer preferences, make data-driven decisions, predict trends, optimize inventory management, and deliver personalized shopping experiences.
- By identifying shoplifting and OTHERS potentially suspicious activity, we improve security for both the supply chain and shoppers.
personal computer
Generative AI
Generative AI (GenAI) is increasingly being integrated into personal computers, revolutionizing how individuals interact with technology.
Thanks to GenAI, personal computers can now generate original content such as images, videos, and music from simple text input.
These AI-powered applications analyze vast amounts of data and learn patterns and styles, enabling them to deliver unique content tailored to users' information and preferences.
Generative AI in personal computers, used for tasks such as creating presentation slides, drafting emails, and summarizing meetings, enhances user productivity and enables self-expression in new and creative ways. As this technology evolves, the possibilities for efficiency and innovation are expected to expand infinitely.
Links
Corporate Profile
| Corporate Name | HAILO TECHNOLOGIES LTD. |
|---|---|
| Headquarters Location | Israel |
| Established | 2017年 |
Inquiry
Columns
Real-time object detection using a fanless edge AI accelerator (YOLOv8m)
This document explains the procedure for running Hailo-8 on a ROCK 5 Model B and building and executing real-time object detection via GStreamer/OpenCV from the HDMI input.
- AI/Generation AI
Edge AI voice assistant with Raspberry Pi 5 and AI HAT+ 2
This article explains the steps and settings to run voice recognition and LLM at high speed on the edge using Raspberry Pi 5 and AI HAT+ 2 (equipped with Hailo-10H).
- AI/Generation AI
- AI Assistant
- AI Agents