Open weibaogang opened 6 years ago
Maybe you can try smaller batch size.
在2018-03-03 21:47:28,邓维建12213033@bjtu.edu.cn写道:
I encount follow error: I0303 21:40:15.840857 18308 net.cpp:380] instance -> instance F0303 21:40:15.842857 18308 cudnn.hpp:106] Check failed: status == CUDNN_STATUS_SUCCESS (3 vs. 0) CUDNN_STATUS_BAD_PARAM Check failure stack trace: @ 0x7fae6e7ec5cd google::LogMessage::Fail() @ 0x7fae6e7ee433 google::LogMessage::SendToLog() @ 0x7fae6e7ec15b google::LogMessage::Flush() @ 0x7fae6e7eee1e google::LogMessageFatal::~LogMessageFatal() @ 0x7fae6ee89e95 caffe::CuDNNConvolutionLayer<>::LayerSetUp() @ 0x7fae6efbfe27 caffe::Net<>::Init() @ 0x7fae6efc256e caffe::Net<>::Net() @ 0x7fae6ef56a55 caffe::Solver<>::InitTrainNet() @ 0x7fae6ef57e45 caffe::Solver<>::Init() @ 0x7fae6ef5815f caffe::Solver<>::Solver() @ 0x7fae6ee06341 caffe::Creator_SGDSolver<>() @ 0x40adb8 train() @ 0x4075b8 main @ 0x7fae6cf82830 __libc_start_main @ 0x407e89 _start @ (nil) (unknown) many people solution this problem with use bigger gpu,so I want know your code need how much gpu memery while train(train_resnet_linear.prototxt);And do you know this problem can be solution with other method?Thank you!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
I encount follow error: I0303 21:40:15.840857 18308 net.cpp:380] instance -> instance F0303 21:40:15.842857 18308 cudnn.hpp:106] Check failed: status == CUDNN_STATUS_SUCCESS (3 vs. 0) CUDNN_STATUS_BAD_PARAM Check failure stack trace: @ 0x7fae6e7ec5cd google::LogMessage::Fail() @ 0x7fae6e7ee433 google::LogMessage::SendToLog() @ 0x7fae6e7ec15b google::LogMessage::Flush() @ 0x7fae6e7eee1e google::LogMessageFatal::~LogMessageFatal() @ 0x7fae6ee89e95 caffe::CuDNNConvolutionLayer<>::LayerSetUp() @ 0x7fae6efbfe27 caffe::Net<>::Init() @ 0x7fae6efc256e caffe::Net<>::Net() @ 0x7fae6ef56a55 caffe::Solver<>::InitTrainNet() @ 0x7fae6ef57e45 caffe::Solver<>::Init() @ 0x7fae6ef5815f caffe::Solver<>::Solver() @ 0x7fae6ee06341 caffe::Creator_SGDSolver<>() @ 0x40adb8 train() @ 0x4075b8 main @ 0x7fae6cf82830 __libc_start_main @ 0x407e89 _start @ (nil) (unknown) many people solution this problem with use bigger gpu,so I want know your code need how much gpu memery while train(train_resnet_linear.prototxt);And do you know this problem can be solution with other method?Thank you!