AI Edge Computing Box

   High-performance quad-core Mali-G610 GPU supports at least two 4K UI streams and handles complex graphics smoothly.

   Expandable 4G, Wi-Fi 6, and Bluetooth modules simplify edge deployments.

   Expandable mSATA SSD storage supports large data-retention requirements.

   Big.LITTLE architecture with four A76 and four A55 cores, CPU frequency up to 2.4 GHz, 1 MB L2 cache, and 3 MB L3 cache.

   One or two AI accelerator modules provide up to 32 TOPS INT8 each and up to 38 TOPS INT8 combined for demanding workloads.

   Supports TensorFlow, Caffe, PyTorch, TFLite, and ONNX, including models for face detection, tracking, recognition, pose estimation, and safety-helmet detection.