I am developing LSTM network in that I am getting error as "Invalid training data. Sequence responses must have the same sequence length as the corresponding predictors"

clc

clear all

close all

XTrain = xlsread('R1_all_data.xlsx',1,'A1:G3788');

YTrain = xlsread('R1_all_data.xlsx',1,'H1:H3788');

XTest = xlsread('R2_all_data.xlsx',1,'A1:G3788');

YTest = xlsread('R2_all_data.xlsx',1,'H1:H3788');

inputSize = 3788;

numResponses = 1;

numHiddenUnits = 1000;

layers = [ sequenceInputLayer(inputSize)

lstmLayer(numHiddenUnits)

fullyConnectedLayer(numResponses)

regressionLayer ];

opts = trainingOptions('adam', 'MaxEpochs', 1000, 'GradientThreshold', 0.01, 'InitialLearnRate', 0.01);

net = trainNetwork(XTrain,YTrain,layers,opts);

YPred1=predict(net,XTest)

figure,

plot(YPred1,'r-*','LineWidth',1),hold on

plot(YTest,'g-*','LineWidth',1), hold all;

xlabel('sample')

ylabel('values')

grid on

legend('Predicted', 'Original ');

title('performance')

Answer 1

[Krishna]

Hi PRAMOD,

It seems there's a misunderstanding in how you're utilizing the ‘sequenceInputLayer.’ The correct number of inputs for the sequence should be 7, not the 3788 you're currently using, which represents the count of your observations. Additionally, it's important to note that both the ‘sequenceInputLayer’ and the subsequent ‘lstmLayer’ are designed to accept only cell input arrays consisting of sequences.

So to give you an perspective they accept data in the format, axbxc where a are the number of sequences you have b is the number of feature in the sequence and c is the number of observation in that sequence.

Also, I noticed that you have not specified ‘sequence’ in lstmLayer. If you are doing sequence to sequence problem which I think you are doing seeing your dataset, please classify ‘sequence’ in ‘lstmLayer’.

For accurate implementation details, please consult the documentation for ‘sequenceInputLayer’ and ‘lstmLayer’,

https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.sequenceinputlayer.htmlhttps://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.lstmlayer.html

Also please go through the following example for detailed explanation on how to use these layers,

https://www.mathworks.com/help/deeplearning/ug/classify-sequence-data-using-lstm-networks.html

Hope this helps.