How can you use CPU with the Parallel Computing Toolbox?
显示 更早的评论
I need to run a Convolutional Neural Networks code, and this requires the PCT but I don't have a NVIDIA GPU. There used to be a parameter in trainingOptions named 'ExecutionEnvironment' where you could choose 'cpu' instead of 'gpu' but it seems to be obsolete.
Is there any other way??
opts = trainingOptions('sgdm', ...
'Momentum', 0.9, ...
'InitialLearnRate', 0.001, ...
'LearnRateSchedule', 'piecewise', ...
'LearnRateDropFactor', 0.1, ...
'LearnRateDropPeriod', 8, ...
'L2Regularization', 0.004, ...
'MaxEpochs', 40, ...
'MiniBatchSize', 128, ...
'Verbose', true);
cifar10Net = trainNetwork(X_train, Labels_train, layers, opts);
采纳的回答
mizuki
2017-5-25
1 个投票
CNN on CPU and 'ExecutionEnvironment' option are introduced in R2017a. If you are using R2016b or previous versions, you need GPU to run it or upgrade MATLAB.
1 个评论
dhikra H
2017-11-22
hello, I use CNN on CPU with matlab 2017b and when i put parallel options in 'ExecutionEnvironment' it display an error indicat that The CUDA runtime failed to initialize correctly. This can occur with some older CUDA drivers. Thanks for help
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 GPU Computing 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
