0

我正在尝试使用交叉熵来实现 RNN。以下是我的代码:

net = layrecnet(1:2,10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
[Xs,Xi,Ai,Ts] = preparets(net, featureMatrix, labels);
net = train(net,Xs,Ts,Xi,Ai);
% view(net)
Y = net(Xs,Xi,Ai);
perf = perform(net,Y,Ts);

来自performParamMatlab 的官方文档。但是,在我执行它之后,我收到一个警告说:

Warning: Performance function replaced with squared
error performance. 
> In trainlm>formatNet (line 155)
  In trainlm (line 65)
  In nntraining.setup (line 14)
  In network/train (line 335) 

即使我执行前馈网络,我也会收到同样的警告。以下是我的代码。

[x,t] = simplefit_dataset;
net = feedforwardnet(10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
net = train(net,x,t);
view(net)
y = net(x);
perf = perform(net,y,t);

那么如何在我的代码中使用交叉熵呢?

4

1 回答 1

0

该问题trainlm仅适用于使用Jacobian Matrix文档中所述的损失函数:

此函数使用雅可比行列式进行计算,假设性能是误差平方的平均值或总和。因此,使用此函数训练的网络必须使用 mse 或 sse 性能函数。

一种解决方案是使用其他训练算法,例如trainrpor trainscg。以下作品:

[x,t] = simplefit_dataset;
net = feedforwardnet(10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
net.trainFcn = 'trainrp';
net = train(net,x,t);
view(net)
y = net(x);
perf = perform(net,y,t);
于 2016-05-12T14:11:02.820 回答