MATLAB: Training data for Autoencoder is limited in size

Question

Dear all,

I have a dataset for training an autoencoder with size of 1×185823 cell, each cell contains a matrix (29×12 double).

However, there will be errors when I train the autoencoder using dataset with the size larger than 1876…This is so weird because the number is so random.

For example,

 autoenc_Data_1_99_Acq = trainAutoencoder(Data_for_Training(1:1876));

This will be fine.

But If the size of training data is large than 1876 cells, there will be errors:

 autoenc_Data_1_99_Acq = trainAutoencoder(Data_for_Training(1:1877));
Error using Autoencoder>iConvertToDouble (line 918)
Training data for an autoencoder must either be a matrix, or a cell array of equal size images.
Error in Autoencoder.train (line 506)
             [X,visualizationDimensions, trainedOnImages] = iConvertToDouble(X);
Error in trainAutoencoder (line 105)
autoenc = Autoencoder.train(X, autonet, paramsStruct.UseGPU);
Error in AutoEncoderTestRawData4 (line 80)
autoenc_Data_1_99_Acq = trainAutoencoder(Data_for_Training(1:1877));

Answer 1

Matrix at 1877 is not 29x12.