Sparse Autoencoder with Adam optimization

6 次查看（过去 30 天）
Amy
Amy 2021-1-5
Hello!
I have a data set that contains 4 parts 1- Train Attribute( 121x125973 double ) , 2- Train Label (1x125973 double ), 3- Test Attribute(121x22544 double ) , 4- Test Label (1x22544 double) for NSL KDD dataset and it is ready to implement algorithem.
I applied sparse autoencoder and works with out any problem
options.Method = 'lbfgs' ;
options.maxIter = maxIter ;
options.useMex = 0 ;
[opttheta, cost] = minFunc( @(p)sparseAutoencoderCost(p, inputSize, ...
hs, l1, sp, beta, trainAttr), theta, options) ;
trainFeatures = feedForwardAutoencoder(opttheta, hs, inputSize, ...
trainAttr) ;
testFeatures = feedForwardAutoencoder(opttheta, hs, inputSize, ...
testAttr) ;
But when I try to optimize the result using Adam optimizer I faced this problem " Unrecognized property 'GRADIENTDECAYFACTOR' for class 'nnet.cnn.TrainingOptionsADAM'.
this is my code
options = trainingOptions('adam', ...
'InitialLearnRate',3e-4, ...
'SquaredGradientDecayFactor',0.99, ...
'MaxEpochs',20, ...
'MiniBatchSize',64, ...
'Plots','training-progress');
[opttheta, cost] = minFunc( @(p)sparseAutoencoderCost(p, inputSize, ...
hs, l1, sp, beta, trainAttr), theta, options) ;
trainFeatures = feedForwardAutoencoder(opttheta, hs, inputSize, ...
trainAttr) ;
testFeatures = feedForwardAutoencoder(opttheta, hs, inputSize, ...
testAttr) ;
I wonder how can apply sparse autoencoder with adam optimization ?

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Eigenvalue Problems 的更多信息

标签

产品


版本

R2020a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by