Can I use cpu instead of gpu for the DeepLearni​ngImageCla​ssificatio​nExample.m

I was running the "DeepLearningImageClassificationExample.m" example but this failed because of unsupported GPU. I did see that this example requires a "CUDA-capable GPU card" but is there a manner to to use the parameters ['ExecutionEnvironment','cpu'] in the "activation" function, and use the CPU instead?

Answer 1

[Joss Knight]

Yes

[Fuad Noman]

How?

[Walter Roberson]

On line 68 change

trainingFeatures = activations(net, augmentedTrainingSet, featureLayer, ...
    'MiniBatchSize', 32, 'OutputAs', 'columns');

to

trainingFeatures = activations(net, augmentedTrainingSet, featureLayer, ...
    'MiniBatchSize', 32, 'OutputAs', 'columns', 'ExecutionEnvironment', 'cpu');