compute frame mean file failed about c3d HOT 2 CLOSED

@TinkerBell123 have you sorted out your issue?

from c3d.

yes,i sorted it out .
But there comes another problem,when i run train_ucf101.sh:
F1202 20:52:52.343057 34102 syncedmem.cpp:47] Check failed: error == cudaSuccess (2 vs. 0) out of memory
*** Check failure stack trace: ***
@ 0x7f3a64a5016d google::LogMessage::Fail()
@ 0x7f3a64a5202f google::LogMessage::SendToLog()
@ 0x7f3a64a4fd03 google::LogMessage::Flush()
@ 0x7f3a64a5294e google::LogMessageFatal::~LogMessageFatal()
@ 0x47582f caffe::SyncedMemory::mutable_gpu_data()
@ 0x45d252 caffe::Blob<>::mutable_gpu_data()
@ 0x4e3f09 caffe::Convolution3DLayer<>::Forward_gpu()
@ 0x47bce8 caffe::Net<>::ForwardPrefilled()
@ 0x465479 caffe::Solver<>::Test()
@ 0x4678d1 caffe::Solver<>::Solve()
@ 0x4374bd main
@ 0x7f3a61567d5d __libc_start_main
@ 0x436e5d (unknown)

I even set both train and test the batch size to 1,there is still out of memory .
can you give some suggestions?

from c3d.