CUDA fails to launch while running seq2seq model
1 view (last 30 days)
Show older comments
When I train the sequence to sequence model. This error occurred--
Warning: An unexpected error occurred during CUDA execution. The CUDA error was: CUDA_ERROR_LAUNCH_FAILED
After that, I can't use the GPU to run any function like gpuArray(). There is no problem when I train any image classification model after restarting the Matlab or before running this model.
model: #sample 1250 in each segment
layers = [ ...
sequenceInputLayer(40)
bilstmLayer(200,'OutputMode','sequence')
dropoutLayer(0.5)
bilstmLayer(150,"OutputMode","sequence")
dropoutLayer(0.5)
bilstmLayer(100,"OutputMode","sequence")
dropoutLayer(0.5)
fullyConnectedLayer(4)
softmaxLayer
classificationLayer];
options = trainingOptions('adam', ...
'MaxEpochs',20, ...
'MiniBatchSize',32, ...
'InitialLearnRate',0.01, ...
'LearnRateDropPeriod',5, ...
'LearnRateSchedule','piecewise', ...
'ValidationData',{fsstValidationData(:,1),labelValidationData},...
'ValidationFrequency',5,...
"ValidationPatience",inf,...
'GradientThreshold',1, ...
'Plots','training-progress',...
'shuffle','every-epoch',...
'Verbose',0,...
'DispatchInBackground',true);
GPU: RTX 2060 super 8 GB;
0 Comments
Answers (0)
See Also
Categories
Find more on GPU Computing 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!