www.gusucode.com > nnet 工具箱 matlab 源码程序 > nnet/nnderivative/+nnMex/grad.m
function [gWB,trainPerf,trainN] = grad(net,data,hints) TEMP = zeros(1,ceil(hints.tempSizeBG/8)*8); numMasks = 1; [gWB,trainPerf,trainN] = nnMex.bg ... (net,data.X,data.Xi,data.Pc,data.Pd,data.Ai,data.T,data.EW,data.trainMask,... data.Q,data.TS,numMasks,hints,TEMP,hints.batch); TEMP = [];