Feature Input layer error MATLAB 2021a

I am running the below code on MATLAB 2021a in Ubuntu 20.04.

numFeatures = 21;

numClasses = 3;

layers = [

featureInputLayer(numFeatures,'Name','input')

fullyConnectedLayer(numClasses, 'Name','fc')

softmaxLayer('Name','sm')

classificationLayer('Name','classification')]

I am not sure why I am getting the below error

Unable to resolve the name nnet.internal.cnn.layer.FeatureInput.

Error in featureInputLayer (line 99)

internalLayer = nnet.internal.cnn.layer.FeatureInput(...

It was working fine in MATLAB2020b. Any help is greatly appreciated.

Answer 1

[Chunru]

featureInputLayer is Introduced in R2020b.

Try:

which featureInputLayer

/MATLAB/toolbox/nnet/cnn/featureInputLayer.m

numFeatures = 21;
numClasses = 3;
layers = [
    featureInputLayer(numFeatures,'Name','input')
    fullyConnectedLayer(numClasses, 'Name','fc')
    softmaxLayer('Name','sm')
    classificationLayer('Name','classification')]

layers = 
  4×1 Layer array with layers:

     1   'input'            Feature Input           21 features
     2   'fc'               Fully Connected         3 fully connected layer
     3   'sm'               Softmax                 softmax
     4   'classification'   Classification Output   crossentropyex

[Dushyant Sahoo]

Running the command which featureInputLayer, I get below as an output

/usr/local/MATLAB/R2021a/toolbox/nnet/cnn/featureInputLayer.m

But when I run the next set of functions, I am getting below error

Unable to resolve the name nnet.internal.cnn.layer.FeatureInput.

Error in featureInputLayer (line 99)

internalLayer = nnet.internal.cnn.layer.FeatureInput(...

[Chunru]

Open up the "/MATLAB/toolbox/nnet/cnn/featureInputLayer.m" and set a break pointer. Then run the command again.

[Dushyant Sahoo]

How does that help? There seems to be a problem in inbuilt function "featureInputLayer" which is a read only file. I am not sure how to correct the problem. Everything appears to work in MATLAB2020b but not in 2021a.

[Chunru]

Running inside the .m file allows you to step through the program and locate where things go wrong. Without your machine, that is the most we can suggest. Next thing you can try is to reinstall matlab.