公开(公告)号：US20200234133A1

公开(公告)日：2020-07-23

申请号：US16842145

申请日：2020-04-07

Applicant: SZ DJI TECHNOLOGY CO., LTD.

Inventor： Sijin LI ,  Kang YANG ,  Manhong LIN ,  Zhao YAN

IPC: G06N3/08 ,  G06N3/04

Abstract: A data fixed-point method, includes: calculating a maximum output value in a first target layer of a neural network for each input sample of a plurality of input samples; selecting at least two of a plurality of maximum output values as fixed-point reference values; determining a reference integer part bit width according to each of the fixed-point reference values; and performing an accuracy test based on a preset output value total bit width and each reference integer part bit width, to determine a reference integer part bit width with a highest accuracy as an integer part bit width used by the first target layer when output values are fixed-pointed.

公开(公告)号：US20200285942A1

公开(公告)日：2020-09-10

申请号：US16884729

申请日：2020-05-27

Applicant: SZ DJI TECHNOLOGY CO., LTD.

Inventor： Zhao YAN ,  Lan DONG ,  Lin CHEN ,  Sijin LI ,  Mingming GAO

IPC: G06N3/04 ,  G06N3/063

Abstract: A method for processing across neural networks includes: when processing a last block of a plurality of blocks of an i-th layer of a first neural network, reading data of a first block of a plurality of blocks of a k-th layer of a second neural network from a memory; and processing the first block of the plurality of blocks of the k-th layer of the second neural network according to the data of the first block of the plurality of blocks of the k-th layer of the second neural network after processing the last block of the plurality of blocks of the i-th layer of the first neural network. 1≤i≤N, N is a number of layers of the first neural network; and 1≤k≤M, M is a number of layers of the second neural network.