Custom neural network - input mismatch

1 次查看(过去 30 天)
I have created a custom neural network with following configurations.
multinet = network;
multinet.numInputs = 44;
multinet.numLayers = 2;
multinet.biasConnect = [1;1];
multinet.layerConnect = [0 0;1 0];
multinet.inputConnect = ipconnect;
multinet.outputConnect = [0 1];
multinet.trainfcn = 'trainbfg';
ipconnect is a 2*44 matrix, in which the first row is all ones and the second is all zeros. My input is a 44*72 matrix (h_tr) and the target variables are in a 1*72 (ytr_t) vector. Whenever I try to invoke the train function, it returns the following error.
[m_net,tr] = train(multinet, h_tr, ytr_t);
Error using network/train (line 340)
Number of inputs does not match net.numInputs.
But when I check with multinet.numInputs it returns 44. Really appreciate any inputs on this. Thanks in advance.

回答(1 个)

Jose Marques
Jose Marques 2017-11-25
Please, verify the order of elements. Maybe you can try to transposte the inputs matrix.

类别

Help CenterFile Exchange 中查找有关 Deep Learning Toolbox 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by