Hailo-8 M.2 AI Accelerator Module Compatible with Raspberry Pi 5 Support Linux/Windows Systems Based On The 26TOPS Hailo-8 AI Processor

Features At A Glance Hailo-8 AI M. 2 modulePowered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor 2.5W typical power consumption Scalable, enabling simultaneous processingof multi-streams & multi-models Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks Supports Linux and Windows Supports the temperature range of -40C to 85C PCIe To M. 2 adapterOnboard power monitoring chip and EEPROM, supports real-time monitoring of device power status for more stable operation Raspberry Pi HAT+ compliant Reserved airflow vent, supports installing cooling fan for better heat dissipation of the AI module to improve performance Immersion gold process design, anti-oxidation and more durableHailo-8 AI M.

2
Module ParametersAI performance: 26 TOPS
Form Factor: M.
2 Key MPower supply: 3.3V
5%Power consumption: 2.5W (Typ. ) 8.65W (Max.
)Interface: PCIe Gen3, 4-laneCertificate: CE, FCC
Class AStorage temperature: -40
~ 85COperating temperature: -40
~ 85COperating humidity: 5% ~ 90%RH (no frosting)
Dimensions: 2280mm with breakable extensions to2242mmand 2260mmEquipped with Hailo-8 AI Accelerator to Step Up Your Edge Product PerformanceThe Hailo-8 M. 2 module is an AI accelerator module for AIapplications, based on the 26 tera-operations per second (TOPS) Hailo-8AI processor with high power efficiency. The M. 2 AI accelerator featuresa full PCIe Gen- 3.0 4-lane interface, delivering unprecedented AIperformance for edge devices. The M. 2 module can be plugged into an existing edge device withM. 2 socket to provide low-power deep neural network inferencing. Leveraging Hailo's comprehensive Dataflow Compiler and its support forstandard AI frameworks, customers can easily port their Neural Networkmodels to the Hailo-8 and introduce high-performance AI products to themarket quickly.