neural network type missmatch issue

2 Ansichten (letzte 30 Tage)
Thomas
Thomas am 29 Nov. 2012
I am trying to import trained neural network from a .mat file
When i select it in the current folder, the box below it interprets it corectly as a network
but when i import it to workspace, through double click or load command it is loaded as a useless structure ( on which i cant perform output = net(input) )
how can i work around it? there should be a way since file preview somehow interprets it correctly
  • is it possible to cast that structure back to network?
  • is it possible to load it properly with loadobj? if so how to do it ?
Im using matlab r2010a and network was probably trained on version 2012b

Akzeptierte Antwort

Walter Roberson
Walter Roberson am 29 Nov. 2012
If I recall correctly, networks have changed in very recent MATLAB, possibly R2012b itself. Networks are OOP objects, which are usually not backwards compatible.
(Note: I have not used networks myself; I just sometimes remember things I have seen "pass by" in Answers)

Weitere Antworten (0)

Kategorien

Mehr zu Deep Learning Toolbox finden Sie in Help Center und File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by