AI Powered Edge Compute with SOPHON BM1688 and Video Processing – Asset Track Pro

AI edge computing device with SOPHON BM1688 SoC, octa-core ARM Cortex-A53, neural network TPU, 8GB RAM and high-performance video encoding/decoding.

SKU: ATPOPE-589 Category:

Description

Features

  • Docker container management technology
  • Large language models
  • Powerful multi-channel video AI performance

 

Technical Specifications

SOC SOPHON BM1688
CPU Octa-core ARM Cortex-A53 at 1.6GHz
TPU SOPHGO Neural Network Acceleration Engine TPU, 32T at INT4 Peak, 16T at INT8 Peak, 4T at FP16/BF16, 0.5T at FP32
Supports Caffe, PyTorch, TensorFlow (Lite), ONNX, MXNet
ISP Supports Time-Division Multiplexing for Video Input from up to 6 Sensors, Max Performance: Dual 8M at 30 SDR, 8M at 30 HDR, 4M at 30 HDR + 8M at 15 SDR
VPU Video Decoding: H.264/H.265 (Max: 1920×1080 at 480FPS, 3840×2160 at 120FPS)
Video Encoding H.264/H.265 (Max: 1920×1080 at 300FPS, 3840×2160 at 75FPS)
Image Codec JPEG/MJPEG Baseline (Max JPEG Resolution: 1080P at 480FPS)
RAM 8GB LPDDR4 (4GB/8GB Optional)