Datastore in Deep Network Designer

Hello everybody!

I have a question about the Deep Network Designer app in Matlab as there is very little information and bibliography about it. I don't know how to create or organize a datastore of numerical values and characteristics of different observations of an industrial process ->[observations x characteristics], the thing is that I don't know how to declare the predictors and responses of the system.

Also in the app designer, I place a featureinputLayer with inputSize the number of predictors (features) that I have, and a fullyConnectedLayer with the number of responses I expect to get.

In the end, I always get the error "Invalid training data. For a network with 1 inputs and 1 output, the datastore read function must return a cell array with 2 columns, but it returns a cell array with 1 columns" or something like that.

I hope I have explained well, if anyone knows more about this app and how to learn I would appreciate it!

Answer 1

[Ajay Gajulapally]

Hi Fernando,

Kindly go through the MATLAB documentation to know more about Deep Network Designer App:

Get Started with Deep Network Designer - MATLAB & Simulink - MathWorks India

Deep Network Designer App - MATLAB & Simulink - MathWorks India

Design, visualize, and train deep learning networks - MATLAB - MathWorks India

[FERNANDO CALVO RODRIGUEZ]

Thank you!

I hope to eventually explain deep learning with matlab in more detail as it has great potential!

[Philip Brown]

There's also this doc page: https://uk.mathworks.com/help/deeplearning/ug/import-data-into-deep-network-designer.html which may help with your questions around data.