use of data store for waveform input in deep learning

I am attempting to input a large amount of waveform data to the deep learning toolbox. It appears that this is best done with data stores. I have not been able to find an example where a non-image datastore is constructed which includes the required label for each waveform. Any example showing the code used to create such a data store would be appreciated.

Answer 1

[Taylor]

I believe what you're looking for is a signalDatastore. Here's an example on using signal datatstores and general documentation.

[Mark Hubelbank]

That looks like a good start. It appears that my problem is that mathlab seems to allow the deep learning engine to figure out what data it should use. There does not seem to be a specific requirement for the format, type, or variable label of the data. This is so foreign to those of us who have used C forever that it is very hard to switch gears.

Am I wrong in thinking that there is no specific requirement for the data and label variable names when creating the .mat file to be used to build the signalDatastore? If there are specific requirements, I have not been able to find them.