I keep getting this error : 'Warning: index in position 1 exceeds array bounds (must not exceed 1)' while running Long short-term memory network

1 view (last 30 days)
load LSTMdata.mat
% LSTM Architecture:
% sequenceInputLayer -> bilstmLayer -> fullyConnectedLayer
% -> softmaxLayer -> classificationLayer
inLayer = sequenceInputLayer(3,'Name','inLayer');
lstm = bilstmLayer(100,'Name','lstm1','OutputMode','last');
fc = fullyConnectedLayer(2,'Name','fullyconnected');
sm = softmaxLayer('Name','softmax');
cf = classificationLayer('Name','classification');
layers = [inLayer;lstm;fc;sm;cf];
options = trainingOptions('adam',...
'Shuffle','every-epoch',...
'Plots','training-progress',...
'LearnRateDropPeriod',200,...
'LearnRateSchedule','piecewise');
net = trainNetwork(xtraindata,ytraindata,layers,options);
testPred = classify(net,XTest)
xtraindata and xtestdata are 140*1 cell array (each cell contains 3*1152 double array) and ytraindata is 140*1 categorical array. I also downloaded a sample neural network from mathworks website but I got the same error when I ran it. Please help!

Answers (1)

Vishal Bhutani
Vishal Bhutani on 10 Sep 2018
Based on the information provided, the network architecture seems to be fine. You can just check one thing that the training and testing data are of same dimension. It would be helpful to provide the .mat file.

Categories

Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange

Products


Release

R2018a

Community Treasure Hunt

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

Start Hunting!