How can I check the existance of sub-variable in network layer

Question

Brian Park 2023-7-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1995288-how-can-i-check-the-existance-of-sub-variable-in-network-layer

评论： Brian Park 2023-7-13

采纳的回答： Angelo Yeo

在 MATLAB Online 中打开

I'm trying to sort the layers which its 'WeightLearnRateFactor' is bigger than 0 for training a network with freezing.

So, I used isfield function to check if there is 'WeightLearnRateFactor' sub-variable in a layer.

existance = isfield(net.Layers(5, 1), 'WeightLearnRateFactor')
>> Result : existance -> 0

but it prints 0 for result, even though there is WeightLearnRateFactor in net.Layer(5,1)...

I tested it with 'Name' and other variables and still can't find sub-variables..

How can I sort the layers with non-zero WeightLearnRateFactor ???

net is dlnetwork type.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Angelo Yeo 2023-7-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1995288-how-can-i-check-the-existance-of-sub-variable-in-network-layer#answer_1272318

在 MATLAB Online 中打开

The 'WeightLearnRateFactor' is a property of Convolution2DLayer not a field. You can use isprop.

net = googlenet;
lgraph = layerGraph(net);
lgraph = removeLayers(lgraph,["prob" "output"]);
dlnet = dlnetwork(lgraph);
isprop(dlnet.Layers(5,1), "WeightLearnRateFactor")
ans = logical
   0
isprop(dlnet.Layers(6,1), "WeightLearnRateFactor")
ans = logical
   1