How to use gpu for deep learning

I’m training detection model yolov4 on matlab. I just got a computer with a graphic card, the Nvidia GeForce RTX 3070 Ti. I want to get the maximum from it. Please help me, what I need to write in matlab code to perform training using GPU.

Answer 1

[KSSV]

Check the trainingOptions, in there you have option to specify the execution environment.

Example:

    options = trainingOptions('sgdm', ...
        'Momentum',0.9, ...
        'InitialLearnRate',initLearningRate, ...
        'LearnRateSchedule','piecewise', ...
        'LearnRateDropPeriod',learningDropPeriod, ...
        'LearnRateDropFactor',learningRateFactor, ...
        'L2Regularization',l2reg, ...
        'ExecutionEnvironment', 'auto',....
        'ValidationPatience',Inf,...
        'MaxEpochs',maxEpochs, ...
        'ValidationData',{inputVal, targetVal}, ...
        'ValidationFrequency',50,...
        'shuffle','every-epoch',....
        'MiniBatchSize',miniBatchSize, ...
        'GradientThresholdMethod','l2norm', ...
        'GradientThreshold',0.01, ...
        'Plots','training-progress', ...
        'ExecutionEnvironment', 'auto',.....  %<------ check this. Keep it auto so MATLAB can pick the best
        'ValidationPatience', 10, ...
        'Verbose',true);