MATLAB Answers

5

How to use a sequenceInputLayer with a regressionLayer (neural networks) ?

Asked by mathbourne on 5 Oct 2017
Latest activity Answered by Greg Heath
on 4 Apr 2019
Hi everyone,
How is it possible to use a sequenceInputLayer with a regressionLayer (neural networks)? I have tried the following code:
layers = [...
sequenceInputLayer(inputSize)
fullyConnectedLayer(30)
fullyConnectedLayer(1)
regressionLayer]
which results with the following error message:
Error using trainNetwork (line 140)
The output of layer 3 is incompatible with the input expected by layer 4.
I searched through the manual and have yet to find something that pertains to the expected input of a regressionLayer.
If someone happens to be able to share the relevant documentation or a snippet of code, that would be really helpful.
Thank you very much!

  2 Comments

same question here. commenting so I get notified.
same here, stuck on this Did you find a solution?

Sign in to comment.

2 Answers

Answer by Greg Heath
on 12 Oct 2017

The regression layer is the only one that has no explicit indication of size.
Hope this helps.
Greg

  2 Comments

Can you put an example code for the solution?
Thanks

Sign in to comment.


Answer by Greg Heath
on 4 Apr 2019

The answer is obvious:
help regressionlayer
doc regressionlayer
*Thank you for formally accepting my answer*
Greg

  0 Comments

Sign in to comment.