Error using trainNetwork. Invalid network.

net = googlenet('Weights','none');
lys = net.Layers;
lys(end-3:end)
numClasses = numel(categories(imdsTrain.Labels));
lgraph = layerGraph(lys);
%Replace the classification layers for new task
newFCLayer = fullyConnectedLayer(3,'Name','new_fc','WeightLearnRateFactor',10,'BiasLearnRateFactor',10);
lgraph = replaceLayer(lgraph,'loss3-classifier',newFCLayer);
newClassLayer = classificationLayer('Name','new_classoutput');
lgraph = replaceLayer(lgraph,'output',newClassLayer);
analyzeNetwork(lgraph)
%Resize the image
imageSize=net.Layers(1).InputSize;
augmentedTrainingSet=augmentedImageDatastore(imageSize,...
    imdsTrain,'ColorPreprocessing','gray2rgb');
augmentedValidateSet=augmentedImageDatastore(imageSize,...
    imdsValidation,'ColorPreprocessing','gray2rgb');
options = trainingOptions('sgdm', ...
    'MiniBatchSize',4, ...
    'MaxEpochs',8, ...
    'InitialLearnRate',1e-4, ...
    'Shuffle','every-epoch', ...
    'ValidationData',augmentedValidateSet, ...
    'ValidationFrequency',3, ...
    'Verbose',false, ...
    'ExecutionEnvironment','cpu', ...
    'Plots','training-progress');
trainedNet = trainNetwork(augmentedTrainingSet,lgraph,options);

ERROR:

Error using trainNetwork

Invalid network.

Caused by:

Layer 'inception_3a-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_3b-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_4a-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_4b-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_4c-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_4d-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_4e-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_5a-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Layer 'inception_5b-output': Unconnected input. Each layer input must be connected to the output of another

layer.

Answer 1

[Sugandhi]

Hi Tan,

I understand that you are facing error using trainNetwork.

The error message indicates that there are unconnected inputs in the network. Specifically, the error message indicates that the input layers of some of the intermediate layers in the network graph are not connected to any output layers. This can commonly occur if you remove layers from the network and do not update the graph accordingly.

To fix the error, you should check the layer graph for any unconnected layers. You can do this using the `analyzeNetwork` function:

analyzeNetwork(lgraph)

This function will display a graph of the layers in your network and highlight any unconnected nodes.

You can then use the `connectLayers` function to connect the unconnected layers to the output of the preceding layers. For example, to connect the `inception_3a-output` layer to the `inception_3a/5x5_reduce` layer, you can use the following code:

lgraph = connectLayers(lgraph,'inception_3a/output','inception_3a/5x5_reduce');

Repeat this step for any other unconnected layers in the network. After you have connected all the layers, you can try training the network again.

Note that if you have removed or modified any layers in the original network, you may need to make additional modifications to the layer graph to ensure that the network is properly connected.

[Tan]

> lgraph = connectLayers(lgraph,'inception_3a-output','inception_3a-5x5_reduce');

Error using nnet.cnn.LayerGraph>iValidateEndLayerInputIsNotOccupied

Unable to connect to 'inception_3a-5x5_reduce'. This input is already connected.

Error in nnet.cnn.LayerGraph/connectLayers (line 310)

iValidateEndLayerInputIsNotOccupied( ..

[Sugandhi]

> lgraph = connectLayers(lgraph,'inception_3a-output','inception_3a-5x5_reduce');

Above code is just an example, you should connect unconnected inputs in the network to any output.The error says that 'inception_3a-5x5_reduce' is already connected.

[Tan]

[Tan]

it shows that inception_3a output. unconnected input.

[Tan]

while i connect it with the previous layer it want me to specify connect to which input

Answer 2

[kalai]

Training with trainNetwork failed.

Invalid network:

Layer 'conv1': Invalid input data. The number of channels of the input data (1) must match the layer's expected number of channels (3)