Implementing Random Seed for Machine Learning

Sofia Brown

2021 9 14

0 个回答

11 次查看（30 天）

0 个投票

I am building a convolutional network and it was suggested to me to set the same random seed at the beginning of my code using the rng command in order to achieve reproducibility of the training results. Here is my neural network:

rootFolder = 'TrainingAll5Sets';
categories = {'0deg', 'eighthdeg'};
rng(0);
%imds = imageDatastore(fullfile(rootFolder, categories), 'LabelSource', 'foldernames');
imds = imageDatastore(fullfile(rootFolder, categories), 'LabelSource', 'foldernames','FileExtensions','.png');
%Define Layers
layers = [
    imageInputLayer([256 320 1])
    convolution2dLayer(1,5,'Padding',2)
    batchNormalizationLayer
    reluLayer
    maxPooling2dLayer(2,'Stride',2)
    convolution2dLayer(6,15,'Padding','same')
    batchNormalizationLayer
    reluLayer
    maxPooling2dLayer(2,'Stride',2)
    convolution2dLayer(12,40,'Padding','same')
    batchNormalizationLayer
    reluLayer
    fullyConnectedLayer(2)
    softmaxLayer
    classificationLayer];
%Set training options - use default options from 7.15.20
options = trainingOptions('sgdm', ...
    'InitialLearnRate',0.00001, ...
    'MaxEpochs',300, ...
    'Shuffle','every-epoch', ...
    'Verbose',false, ...
    'Plots','training-progress');
%Train
[net, info] = trainNetwork(imds, layers, options);

My question is have I implemented the rng command correctly in line 3? My understanding is that rng will generate the same random weights so that I will get reproducible results. Is inserting rng in the beginning of my code sufficient to do this? I want to understand how rng works in a machine learning algorithm.