errors: Point Cloud Classification Using PointNet Deep Learning

'modelGradients' :

Generate Synthetic Signals Using Conditional GAN

Model-Based Reinforcement Learning Using Custom Training Loop

error: deep.internal.dlfeval (Line 17)

[varargout{1:nargout}] = fun(x{:});

error: deep.internal.dlfevalWithNestingCheck (Line 19)

[varargout{1:nargout}] = deep.internal.dlfeval(fun,varargin{:});

error: dlfeval (Line 31)

[varargout{1:nargout}] = deep.internal.dlfevalWithNestingCheck(fun,varargin{:});

error:

[gradients, loss, state, acc] = dlfeval(@modelGradients,XTrain,YTrain,parameters,state);

[Walter Roberson]

I would expect the error message to indicate what exactly MATLAB thinks is going wrong.

You are probably going to need to

dbstop if error

and run the code.

Answer 1

[Samay Sagar]

Hi Lu,

I see that you are encountering the above error while running the example “Generate Synthetic Signals Using Conditional GAN“ available in the MathWorks documentation.

You can open the example files by executing the following command in the MATLAB command window:

openExample('deeplearning_shared/GenerateSyntheticPumpSignalsUsingCGANExample')

The “modelGradient” function expects these arguments: dlnetGenerator, dlnetDiscriminator, dlX, dlT, dlZ

You can check the implementation of the same by executing the following command in the MATLAB command window:

openExample('deeplearning_shared/GenerateSyntheticPumpSignalsUsingCGANExample','supportingFile','modelGradients.m')

When you execute the code blocks under the “Train model” section, you can execute the following command to use the “modelGradient” function:

dlfeval(@modelGradients, dlnetGenerator, dlnetDiscriminator, dlXGeneratedNew, dlTNew, dlZNew)

Make sure that the variables have been evaluated before executing the above function.Hope this helps!

Hope this helps!