I do not know if it is because I used the normalization layer. My training data volume is about 8T, since TrainedMean and TrainedVariance are calculated after the training is finished,so this process costs a lot of time.
Is it normal to stop and save a network for more than 14 hours?
1 次查看(过去 30 天)
显示 更早的评论
I saved my network using the following function when it completed training progress(Fig.1). But after more than 14 hours, the network still has not finished saving process. The GPU has always been running during the saving process (Fig.2). The training data is about 8T.
net = trainNetwork(trainData,lgraph,options);
save([dir,'Anet'],'net','-v7.3')
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/434408/image.png)
Fig. 1
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/434413/image.png)
Fig. 2
0 个评论
回答(1 个)
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Image Data Workflows 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!