why this Warning message occur: Support for GPU devices with Compute Capability 5.0 will be removed in a future MATLAB release? how can fix it?

I have a matlab code to classify x-ray dataset. I used the Googlenet model with augmentation for classiying. But each time I try to run the code a warning message occur.

the matlab code is:

clear

clc

datapath='dataset';

imds=imageDatastore(datapath,'includeSubfolders',true,'LabelSource','foldernames');

total_split=countEachLabel(imds);

num_images=length(imds.Labels);

perm=randperm(num_images,4);

figure;

for idx=1:length(perm)

subplot(2,2,idx);

imshow(imread(imds.Files{perm(idx)}));

title(sprintf('%s',imds.Labels(perm(idx))))

end

num_folds=5;

for fold_idx=1:num_folds

fprintf('processing %d among %d folds \n',fold_idx,num_folds);

test_idx=fold_idx:num_folds:num_images;

imdsTest = subset(imds,test_idx);

train_idx=setdiff(1:length(imds.Files),test_idx);

imdsTrain = subset(imds,train_idx);

net=googlenet;

lgraph=layerGraph(net);

clear net;

numClasses = numel(categories(imdsTrain.Labels));

newLearnableLayer=fullyConnectedLayer(numClasses,...

'name','new_fc',...

'weightLearnRateFactor',2,...

'BiasLearnRateFactor',2);

lgraph=replaceLayer(lgraph,'loss3-classifier',newLearnableLayer);

newsoftmaxLayer = softmaxLayer('Name','new_softmax');

lgraph = replaceLayer(lgraph,'prob',newsoftmaxLayer);

newClassLayer=classificationLayer('Name','new_classoutput');

lgraph = replaceLayer(lgraph,'output',newClassLayer);

imdsTrain.ReadFcn=@(filename)preprocess_xray(filename);

imdsTest.ReadFcn=@(filename)preprocess_xray(filename);

options=trainingOptions('adam',...

'MaxEpochs',10,'miniBatchSize',8,...

'Shuffle','every-epoch',...

'InitialLearnRate',1e-4,...

'Verbose',false,...

'Plots','training-progress');

augmenter = imageDataAugmenter( ...

'RandRotation',[-5 5],'RandXReflection',1,...

'RandYReflection',1,'RandXShear',[-0.05 0.05],'RandYShear',[-0.05 0.05]);

auimds = augmentedImageDatastore([224 224],imdsTrain,'DataAugmentation',augmenter);

netTransfer=trainNetwork(auimds,lgraph,options);

augtestimds=augmentedImageDatastore([224 224],imdsTest);

[predicted_labels(test_idx),posterior(test_idx,:)] = classify(netTransfer,augtestimds);

save(sprintf('ResNet50_%d_among_%d_folds',fold_idx,num_folds),'netTransfer','test_idx','train_idx');

clearvars -except fold_idx num_folds num_images predicted_labels posterior imds netTransfer;

end

when run code, these messages occur:

processing 1 among 5 folds

Warning: Support for GPU devices with Compute Capability 5.0 will be removed in a future MATLAB release. For more information on GPU

support, see GPU Support by Release.

> In parallel.internal.gpu.selectDevice

In parallel.gpu/GPUDevice/current (line 44)

In gpuDevice (line 23)

In nnet.internal.cnn.util.isGPUCompatible (line 10)

In nnet.internal.cnn.util.GPUShouldBeUsed>iCheckLocalGPU (line 51)

In nnet.internal.cnn.util.GPUShouldBeUsed>iValidateGPU (line 85)

In nnet.internal.cnn.util.GPUShouldBeUsed (line 34)

In nnet.internal.cnn.assembler.setupExecutionEnvironment (line 27)

In nnet.internal.cnn.trainNetwork.doTrainNetwork (line 22)

In trainNetwork (line 182)

In try45 (line 46)

processing 2 among 5 folds

processing 3 among 5 folds

processing 4 among 5 folds

What shall I do to fix the problem?

Answer 1

[Joss Knight]

warning("off","parallel:gpu:device:DeviceDeprecated");

Answer 2

[Walter Roberson]

You have several options:

• you can upgrade your gpu hardware to Pascal architecture or later
• you can wait a couple of releases until Maxwell and earlier are not supported at all, in which case the code will stop working with a different error message, solving the problem of receiving this error message
• you could ignore the warning message and deal with the situation when eventually your code fails in a new release
• you could turn off the warning and deal with the situation eventually when your code fails in a new release