Training Semantic Segmentation out of memory

1 次查看（过去 30 天）

evianita dewi 2019-12-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/498099-training-semantic-segmentation-out-of-memory

i have error when trying to training my datasets in semantic segmentation using deep learning. this is my code:

doTraining = true;
if doTraining
    [net,info] = trainNetwork(pximds,lgraph,options);
    save ('D:\TAEvianita\video\tilesdatase\Pretraineddata.mat','net');
    disp('NN trained')
else
    %load the pre-trained network
     data = load('D:\TAEvianita\video\tilesdatase\net_checkpoint__266__2019_12_23__23_44_28.mat')
     net = data.net
end

and this is my training options

options = trainingOptions('sgdm', ...
    'ExecutionEnvironment','gpu',...
    'LearnRateSchedule','piecewise',...
    'LearnRateDropPeriod',10,...
    'LearnRateDropFactor',0.3,...
    'Momentum',0.9, ...
    'InitialLearnRate',1e-3, ...
    'L2Regularization',0.005, ...
    'MaxEpochs',30, ...  
    'MiniBatchSize',2, ...
    'Shuffle','every-epoch', ...
    'CheckpointPath','D:\TAEvianita\video\tilesdatase', ...
    'VerboseFrequency',2,...
    'Plots','training-progress',...
    'ValidationPatience', 4