The Baidu Brain EdgeBoard computing box/card (FZ5) is a deep learning computing box/card built on the Xilinx Zynq Ultrascale CZU5EV chip. It has features such as multi-core processing capability, FPGA programmability, and hardware decoding capability for video streams. It is equipped with a customized deep learning soft core based on the Linux operating system and Baidu's deep learning platform - Paddle, which enables efficient and fast implementation of a series of processes such as model training, deployment, and inference.
The MYIR FZ5 is a deep learning computing box and card based on the Xilinx Zynq Ultrascale CZU 5EV chip, featuring a 4-core ARM A53 processor, GPU, and FPGA. It comes with a customized Linux system and the PaddlePaddle deep learning soft kernel, compatible with Baidu Brain resources and tools like EasyDL and AIStudio.
The document provides instructions on starting the device, connecting via SSH or serial port, and debugging network configurations. It also details how to run model inference examples, including ResNet50 and MobileNet, and explains the advanced development process for creating applications using the prediction library. This manual is written in Chinese.
| Processor | 4-core ARM A53 |
|---|---|
| Default IP Address | 192.168.1.254 |
| Operating System | Linux |
| Data Format | NHWC |
The default username and password are both 'root'.
The default static IP address is 192.168.1.254.