Error in using trainNetwork with trainingOptions verbose as 'true'
4 views (last 30 days)
Show older comments
Im trying to perform image classification on AlexNet. The network trains just fine. However, if I set the verbose to 'true' under the trainingOptions, I get an error as shown below. I cant find the source of the error. Any idea how to fix this error? or is there any other way I can save my training progress (accuracy or loss in every epochs) in a text file or some sort? Thank you.
options = trainingOptions('sgdm', ...
'MiniBatchSize',7, ...
'MaxEpochs',100, ...
'InitialLearnRate',1e-4, ...
'Shuffle','every-epoch', ...
'ValidationData',imdsValidation, ...
'ValidationFrequency',3, ...
'Plots','training-progress',...
'Verbose',false, ...
'OutputFcn',@(info)stopIfAccuracyNotImproving(info,20));
imageAugmenter = imageDataAugmenter( ...
'RandXReflection', true,...
'RandYReflection', true);
augImds = augmentedImageDatastore(inputSize,imdsTrain, ...
'DataAugmentation',imageAugmenter);
net = trainNetwork(augImds,layers,options);
Error using trainNetwork (line 165)
Invalid separator input: string
Error in Alexnet (line 134)
net = trainNetwork(augImds,layers,options);
Caused by:
Error using strjoin (line 37)
Invalid separator input: string
1 Comment
Dheeraj Singh
on 9 Sep 2019
Can you provide the mat file of the variables, to reproduce the problem??
Answers (0)
See Also
Categories
Find more on Image Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!