为什么我的验证设置为空（matlab-patternnet）？

Question

我在自定义输入数据上使用简单的演示代码。 我遇到了奇怪的问题，我的验证集是空的，并且只监视测试集的性能（可能是第一个问题的第二个问题结果）。

>> size(x)

ans =

    25   764

>> size(t)

ans =

     7   764

 trainFcn = 'trainbr';  % Scaled conjugate gradient backpropagation.

% Create a Pattern Recognition Network
hiddenLayerSize = 100;
net = patternnet(hiddenLayerSize, trainFcn);

% Setup Division of Data for Training, Validation, Testing
net.divideParam.trainRatio = 60/100;
net.divideParam.valRatio = 20/100;
net.divideParam.testRatio = 20/100;

% Train the Network
[net,tr] = train(net,x,t);

>> tr

tr = 

  struct with fields:
    
   divideFcn: 'dividerand'
  divideMode: 'sample'
 divideParam: [1×1 struct]
    trainInd: [1×611 double]
      valInd: []
     testInd: [1×153 double]
    
   best_perf: 0.0015
  best_vperf: NaN
  best_tperf: 0.0942

Answer 1

似乎它依赖于方法。其他方法（scg）具有非空验证......