how improove a deep network for image classification

Walid

2026 3 18

0 个回答

30 次查看（30 天）

0 个投票

i used a simple network that gives an accuracy =0.35, I tried to increase the max iteration without improovement.:

numClasses = 5;
layers = [
    imageInputLayer(inputSize)
    convolution2dLayer(3,10)
    batchNormalizationLayer
    reluLayer
    maxPooling2dLayer(2,'Stride',2)
    fullyConnectedLayer(numClasses)
    softmaxLayer];
%%%%%%
options = trainingOptions("sgdm", ...
    'InitialLearnRate', 0.001, ...
    'MaxEpochs', 4, ...         % Increase this value to increase total iterations
    'MiniBatchSize', 32, ...
    ValidationData=imdsValidation, ...
    ValidationFrequency=30, ...
    Plots="training-progress", ...
    Metrics="accuracy", ...
    Verbose=false);