Extract Data for Training

In the workspace, unzip the data.

unzip('MerchData.zip');

Select a Pretrained Network

Open Deep Network Designer.

deepNetworkDesigner

Load a pretrained GoogLeNet network by selecting it from the Deep Network Designer Start Page. If you need to download the network, then clickInstallto open the Add-On Explorer.

Deep Network Designer displays a zoomed-out view of the whole network. Explore the network plot. To zoom in with the mouse, useCtrl+scroll wheel.

Load Data Set

To load the data into Deep Network Designer, on theDatatab, clickImport Data>Import Image Data. The Import Image Data dialog box opens.

In theData sourcelist, selectFolder. ClickBrowseand select the extracted MerchData folder.

The dialog box also allows you to split the validation data from within the app. Divide the data into 70% training data and 30% validation data.

Specify augmentation operations to perform on the training images. For this example, apply a random reflection in the x-axis, a random rotation from the range [-90,90] degrees, and a random rescaling from the range [1,2].

ClickImportto import the data into Deep Network Designer.

Using Deep Network Designer, you can visually inspect the distribution of the training and validation data in theDatatab. You can see that, in this example, there are five classes in the data set. You can also view random observations from each class.

Deep Network Designer resizes the images during training to match the network input size. To view the network input size, in theDesignertab, click theimageInputLayer. This network has an input size of 224-by-224.

Edit Network for Transfer Learning

To retrain a pretrained network to classify new images, replace the last learnable layer and the final classification layer with new layers adapted to the new data set. In GoogLeNet, these layers have the names'loss3-classifier'and'output', respectively.

In theDesignertab, drag a newfullyConnectedLayerfrom theLayer Libraryonto the canvas. SetOutputSizeto the number of classes in the new data, in this example,5.

Edit learning rates to learn faster in the new layers than in the transferred layers. SetWeightLearnRateFactorandBiasLearnRateFactorto10. Delete the last fully connected layer and connect your new layer instead.

Replace the output layer. Scroll to the end of theLayer Libraryand drag a newclassificationLayeronto the canvas. Delete the originaloutputlayer and connect your new layer instead.

Check Network

Check your network by clickingAnalyze. The network is ready for training if Deep Learning Network Analyzer reports zero errors.

Train Network

To train the network with the default settings, on theTrainingtab, clickTrain.

If you want greater control over the training, clickTraining Options并选择设置训练。默认的training options are better suited for large data sets. For small data sets, use smaller values for the mini-batch size and the validation frequency. For more information on selecting training options, seetrainingOptions.

For this example, setInitialLearnRateto0.0001,ValidationFrequencyto5, andMaxEpochsto8. As there are 55 observations, setMiniBatchSizeto11to divide the training data evenly and ensure the whole training set is used during each epoch.

To train the network with the specified training options, clickCloseand then clickTrain.

Deep Network Designer allows you to visualize and monitor the training progress. You can then edit the training options and retrain the network, if required.

Export Results from Training

To export the results from training, on theTrainingtab, selectExport>Export Trained Network and Results. Deep Network Designer exports the trained network as the variabletrainedNetwork_1and the training info as the variabletrainInfoStruct_1.

You can also generate MATLAB code, which recreates the network and the training options used. On theTrainingtab, selectExport>Generate Code for Training.

Test Trained Network

Select a new image to classify using the trained network.

I = imread("MerchDataTest.jpg");

Resize the test image to match the network input size.

I = imresize(I, [224 224]);

Classify the test image using the trained network.

[YPred,probs] = classify(trainedNetwork_1,I); imshow(I) label = YPred; title(string(label) +", "+ num2str(100*max(probs),3) +"%");

For more information, including on other pretrained networks, seeDeep Network Designer.