A problem about "embedAndReshapeLayer"

My matlab version is: '9.13.0.2399474 (R2022b) Update 7'

I try to copy a code section from the official example about the conditional GAN with the link below (the code section is listed behind):

https://www.mathworks.com/help/deeplearning/ug/generate-synthetic-pump-signals-using-conditional-generative-adversarial-network.html?searchHighlight=GAN&s_tid=srchtitle_support_results_23_GAN

Two problems occurred:

1. Seems to be a grammar issue related to layer "projectAndReshapeLayer" definition. The command window indicates:

'Error using projectAndReshapeLayer

projectAndReshapeLayer(projectionSize,numLatentInputs,'proj');

↑

Invalid argument at position 2. A name is expected.

Error in untitled (line 11)

projectAndReshapeLayer(projectionSize,numLatentInputs,'proj');

2. I tried to adjust the above issued statement by keeping projectionSize but deleting numLatentInputs. This statement is all good but another error occured in the command window:

'embedAndReshapeLayer' is used in Generate Synthetic Signals Using Conditional GAN.

Error in untitled (line 32)

embedAndReshapeLayer(projectionSize(1:2),embeddingDimension,numClasses,'emb')];

Do I have to construct the whole GAN (Generator and Discriminator) to solve this error? Instead, I just want to loop more in depth about the data size variation in the generator network.

% Generator Network

numFilters = 64;
numLatentInputs = 100;
projectionSize = [4 1 1024];
numClasses = 2;
embeddingDimension = 100;

layersGenerator = [
    imageInputLayer([1 1 numLatentInputs],'Normalization','none','Name','in')
    projectAndReshapeLayer(projectionSize,numLatentInputs,'proj');
    concatenationLayer(3,2,'Name','cat');
    transposedConv2dLayer([5 1],8*numFilters,'Name','tconv1')
    batchNormalizationLayer('Name','bn1','Epsilon',5e-5)
    reluLayer('Name','relu1')
    transposedConv2dLayer([10 1],4*numFilters,'Stride',4,'Cropping',[1 0],'Name','tconv2')
    batchNormalizationLayer('Name','bn2','Epsilon',5e-5)
    reluLayer('Name','relu2')
    transposedConv2dLayer([12 1],2*numFilters,'Stride',4,'Cropping',[1 0],'Name','tconv3')
    batchNormalizationLayer('Name','bn3','Epsilon',5e-5)
    reluLayer('Name','relu3')
    transposedConv2dLayer([5 1],numFilters,'Stride',4,'Cropping',[1 0],'Name','tconv4')
    batchNormalizationLayer('Name','bn4','Epsilon',5e-5)
    reluLayer('Name','relu4')
    transposedConv2dLayer([7 1],1,'Stride',2,'Cropping',[1 0],'Name','tconv5')
    ];

'projectAndReshapeLayer' is used in the following examples:
Generate Synthetic Signals Using Conditional GAN
Train Variational Autoencoder (VAE) to Generate Images
Include Custom Layer in Network
Train Generative Adversarial Network (GAN)
Train Wasserstein GAN with Gradient Penalty (WGAN-GP)

lgraphGenerator = layerGraph(layersGenerator);

layers = [
    imageInputLayer([1 1],'Name','labels','Normalization','none')
    embedAndReshapeLayer(projectionSize(1:2),embeddingDimension,numClasses,'emb')];

lgraphGenerator = addLayers(lgraphGenerator,layers);
lgraphGenerator = connectLayers(lgraphGenerator,'emb','cat/in2');

Answer 1

[Yang Liu]

I know how to solve the problem now. Just copy the folder "GenerateSyntheticPumpSignalsUsingCGANExample" from the path:

\MATLAB\Examples\R2023b\deeplearning_shared\GenerateSyntheticPumpSignalsUsingCGANExample

to the current working directory and add this folder to path.

One can also just add the related function to path, but in case there may be more to use, copy the whole folder and add it to path will be more time-saving.