Output data size does not match net.outputs{3}.size.

Question

Amanuel Kachiko Kuno 2024-3-7

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2091806-output-data-size-does-not-match-net-outputs-3-size

回答： Ayush Aniket 2024-6-18

I = out.Input;

T = out.Output;

net=newff(minmax(I),[3,5,3],{'logsig', 'tansig','purelin'},'trainlm');

But what i get the error is :

Error using network/train (line 340)

Output data size does not match net.outputs{3}.size

Error in ANN (line 15)

net=train(net,I,T);% Start trining the network

But my inputs are:

out= 1x1SimulationOutput with in that

Input = 4000001x3

Output = 400000x3

I tried a lot Please help me as much as possible. thanks in advance for your guidance.

2 个评论
显示无隐藏无

Alexander 2024-3-7

Can you supply some code that throws this error?

Amanuel Kachiko Kuno 2024-3-7

I = out.Input;

T = out.Output;

net=newff(minmax(I),[3,5,3],{'logsig', 'tansig','purelin'},'trainlm');

net.trainParam.show=1;

net.trainParam.epochs = 1000;

net.trainParam.goal =1e-12;

net=train(net,I,T);

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Ayush Aniket 2024-6-18

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2091806-output-data-size-does-not-match-net-outputs-3-size#answer_1473676

在 MATLAB Online 中打开

Hi Amanuel,

The error you encounter is due to the incorrect syntax of the inputs and targets to the neural network. The network expects the inputs in the form of RxQ1 matrix of Q1 (number of samples) representative R-element input vectors (number of features) and outputs as SNxQ2 matrix of Q2 representative SN-element target vectors. The correct method would be to transpose your I and T matrices as shown below:

%Considering I and T to be in the shape (400000 x 3)
I = out.Input;
T = out.Output;
net=newff(minmax(I'),[3,5,3],{'logsig', 'tansig','purelin'},'trainlm');
net.trainParam.show=1;
net.trainParam.epochs = 1000;
net.trainParam.goal =1e-12; 
net=train(net,I',T');

You can access the documentation for newff function by using the command below:

help newff

Note - The newff function has been obsolete since a long time. You should use feedforwardnetwork function instead, and can read more about it at the following documentation link:

https://www.mathworks.com/help/deeplearning/ref/feedforwardnet.html