Hailo-8L edge AI processing
The Hailo-8L Entry-Level AI Accelerator brings neural network inference to compact, power-conscious edge devices. Part number HNC1LBI11BH is the Hailo-8L processor IC for designs that need up to 13 TOPS of AI compute. Hailo specifies 1.5 W typical processor power consumption and an industrial operating temperature range of −40°C to 85°C.
Designed for embedded computer vision, the Hailo-8L can run AI tasks close to the camera or sensor. Applications include object detection, image classification, video analytics and visual inspection. Its architecture supports concurrent models and multiple video streams, allowing developers to build AI pipelines around their application’s performance and power requirements.
Hardware integration and software
The processor has integrated memory and requires no external DRAM. This can simplify a chip-on-board design and reduce the components needed around the accelerator. It is an AI accelerator IC; customers looking for a ready-to-install M.2 card should use the separate Hailo-8L M.2 module listing.
Hailo’s software suite includes a dataflow compiler for preparing neural network models for deployment. It supports workflows using TensorFlow, TensorFlow Lite, Keras, PyTorch and ONNX, with x86 and Arm host architectures and Linux or Windows environments. Developers can also use the Hailo-8 software ecosystem when planning a move to a higher-capacity Hailo-8 design.
Typical applications
- Smart cameras and on-device video analytics
- Security and access-control systems
- Retail analytics
- Industrial visual inspection
- Embedded object detection and image classification
