faster_rcnn
faster_rcnn copied to clipboard
ShaoqingRen
MATLAB crash when running 'script_faster_rcnn_demo.m' with trained model.(Not pre-trained model)
It runs well when running 'script_faster_rcnn_demo.m' with pre-trained ZF model, which can be gotten from the 'fetch_faster_rcnn_final_model.m'
But MATLAB collapses when I make it running with trained model, which is trained with VOC2007 datasets.
Here's the MATLAB system error message.
MATLAB crash file:C:\Users\Heesu\AppData\Local\Temp\matlab_crash_dump.6600-1:
abort() detected at Mon Jan 16 12:52:23 2017
Configuration: Crash Decoding : Disabled - No sandbox or build area path Crash Mode : continue (default) Current Graphics Driver: Unknown hardware Default Encoding : windows-1252 Deployed : false Graphics card 1 : NVIDIA ( 0x10de ) NVIDIA GeForce GTX 960M Version 21.21.13.6930 Graphics card 2 : Intel Corporation ( 0x8086 ) Intel(R) HD Graphics 530 Version 20.19.15.4539 Host Name : Heesu_BM MATLAB Architecture : win64 MATLAB Entitlement ID: 2121656 MATLAB Root : C:\Program Files\MATLAB\R2016b MATLAB Version : 9.1.0.441655 (R2016b) OpenGL : hardware Operating System : Microsoft Windows 8.1 Pro Processor ID : x86 Family 6 Model 94 Stepping 3, GenuineIntel Virtual Machine : Java 1.7.0_60-b19 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode Window System : Version 6.3 (Build 9600)
Fault Count: 1
Abnormal termination: abort()
Register State (captured): RAX = 000000000455f2e8 RBX = 000000000455f2e8 RCX = 0000000004038560 RDX = 0000000000000000 RSP = 00000000040384e0 RBP = 0000000004039470 RSI = 000000000455f2e8 RDI = 0000000000000000
R8 = 0000000000000000 R9 = 00007ffba2a00000 R10 = 000000000454f287 R11 = 000000000454f287 R12 = 0000000000000000 R13 = 0000000004038ba0 R14 = 000000000454f218 R15 = 000000000455f268
RIP = 000000000438971a EFL = 00000202
CS = 0033 FS = 0053 GS = 002b
Stack Trace (captured): [ 0] 0x0000000004384153 C:\Program Files\MATLAB\R2016b\bin\win64\libmwfl.dll+00082259 fl::diag::thread_context::unspecified_bool+00000051 [ 1] 0x0000000004382a68 C:\Program Files\MATLAB\R2016b\bin\win64\libmwfl.dll+00076392 fl::diag::stacktrace_base::capture+00000024 [ 2] 0x0000000004385b2a C:\Program Files\MATLAB\R2016b\bin\win64\libmwfl.dll+00088874 fl::diag::thread_context::unspecified_bool+00006666 [ 3] 0x00000000043884b7 C:\Program Files\MATLAB\R2016b\bin\win64\libmwfl.dll+00099511 fl::test::terminate_handler::~terminate_handler+00003207 [ 4] 0x0000000007fedcf5 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00580853 mnShutdownMCR+00022277 [ 5] 0x0000000007fed9f8 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00580088 mnShutdownMCR+00021512 [ 6] 0x0000000007fedb0a C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00580362 mnShutdownMCR+00021786 [ 7] 0x0000000007feafeb C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00569323 mnShutdownMCR+00010747 [ 8] 0x0000000007fec4b9 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00574649 mnShutdownMCR+00016073 [ 9] 0x00007ffba2a6ee1d C:\Program Files\MATLAB\R2016b\bin\win64\MSVCR120.dll+00454173 raise+00000489 [ 10] 0x00007ffba2a74a14 C:\Program Files\MATLAB\R2016b\bin\win64\MSVCR120.dll+00477716 abort+00000024 [ 11] 0x00007ffba2215cda E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\libglog.dll+00023770 google::LogMessage::RecordCrashReason+00000058 [ 12] 0x00007ffba2215c6c E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\libglog.dll+00023660 google::LogMessage::SendToLog+00000828 [ 13] 0x00007ffba2215785 E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\libglog.dll+00022405 google::LogMessage::Flush+00000197 [ 14] 0x00007ffba221785b E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\libglog.dll+00030811 google::LogMessageFatal::~LogMessageFatal+00000027 [ 15] 0x00007ffb7dbb9624 E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\caffe_.mexw64+07837220 mexFunction+00110692 [ 16] 0x00007ffb7dc1047c E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\caffe_.mexw64+08193148 mexFunction+00466620 [ 17] 0x00007ffb7db9e89a E:\MATLAB\RCNN\MyRCNN\faster_rcnn-master_BM_170113_1841_TrnOrgDtst\external\caffe\matlab\caffe_faster_rcnn\caffe_.mexw64+07727258 mexFunction+00000730 [ 18] 0x00000000fc5fc5d1 C:\Program Files\MATLAB\R2016b\bin\win64\libmex.dll+00116177 mexRunMexFile+00000129 [ 19] 0x00000000fc5fb2e2 C:\Program Files\MATLAB\R2016b\bin\win64\libmex.dll+00111330 mexFeature_mexver+00001794 [ 20] 0x00000000fc5fa506 C:\Program Files\MATLAB\R2016b\bin\win64\libmex.dll+00107782 mexUnlock+00020182 [ 21] 0x00000000080946de C:\Program Files\MATLAB\R2016b\bin\win64\m_dispatcher.dll+00018142 Mfh_file::dispatch_fh_impl+00000434 [ 22] 0x00000000080944b6 C:\Program Files\MATLAB\R2016b\bin\win64\m_dispatcher.dll+00017590 Mfh_file::dispatch_fh+00000062 [ 23] 0x0000000008095827 C:\Program Files\MATLAB\R2016b\bin\win64\m_dispatcher.dll+00022567 Mfunction_handle::dispatch+00001015 [ 24] 0x0000000008d3bfbd C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00901053 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00811977 [ 25] 0x0000000008d0fbc0 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00719808 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00630732 [ 26] 0x0000000008d38707 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00886535 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00797459 [ 27] 0x0000000008d374b8 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00881848 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00792772 [ 28] 0x0000000008cb9eab C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00368299 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00279223 [ 29] 0x0000000008cb9740 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00366400 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00277324 [ 30] 0x0000000008cb3f8f C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00343951 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00254875 [ 31] 0x0000000008cb3aa1 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00342689 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00253613 [ 32] 0x0000000008cb3991 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00342417 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00253341 [ 33] 0x0000000008e403e9 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+01967081 boost::serialization::singleton<boost::archive::detail::oserializerboost::archive::polymorphic_oarchive,foundation::msg_svc::eventmgr::BaseEvent >::get_instance+00003353 [ 34] 0x0000000008e40372 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+01966962 boost::serialization::singleton<boost::archive::detail::oserializerboost::archive::polymorphic_oarchive,foundation::msg_svc::eventmgr::BaseEvent >::get_instance+00003234 [ 35] 0x0000000008d02a61 C:\Program Files\MATLAB\R2016b\bin\win64\m_lxe.dll+00666209 boost::serialization::singleton<boost::archive::detail::iserializerboost::archive::binaryTerm_iarchive,ir::IrTree >::get_instance+00577133 [ 36] 0x00000000082fb3bd C:\Program Files\MATLAB\R2016b\bin\win64\m_interpreter.dll+00504765 inEvalCmdWithLocalReturn+00000065 [ 37] 0x00000000fb608ced C:\Program Files\MATLAB\R2016b\bin\win64\libmwbridge.dll+00101613 mnGetPrompt+00012957 [ 38] 0x00000000fb609ba3 C:\Program Files\MATLAB\R2016b\bin\win64\libmwbridge.dll+00105379 mnParser+00001091 [ 39] 0x0000000007f96841 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00223297 mcr::runtime::setInterpreterThreadToCurrent+00025585 [ 40] 0x0000000007f957d7 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00219095 mcr::runtime::setInterpreterThreadToCurrent+00021383 [ 41] 0x0000000007f95853 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00219219 mcr::runtime::setInterpreterThreadToCurrent+00021507 [ 42] 0x0000000007f96141 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00221505 mcr::runtime::setInterpreterThreadToCurrent+00023793 [ 43] 0x00000000fd01d347 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00578375 iqm::UserEvalPlugin::pre+00030679 [ 44] 0x00000000fd029e7c C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00630396 iqm::UserEvalPlugin::pre+00082700 [ 45] 0x00000000fd017480 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00554112 iqm::UserEvalPlugin::pre+00006416 [ 46] 0x00000000fd02c99a C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00641434 iqm::UserEvalPlugin::pre+00093738 [ 47] 0x00000000fcff9a37 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00432695 iqm::PackagedTaskPlugin::PackagedTaskPlugin+00000727 [ 48] 0x00000000fcffa08f C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00434319 iqm::PackagedTaskPlugin::execute+00000575 [ 49] 0x00000000fcff9aa9 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00432809 iqm::PackagedTaskPlugin::PackagedTaskPlugin+00000841 [ 50] 0x00000000fcff9f04 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00433924 iqm::PackagedTaskPlugin::execute+00000180 [ 51] 0x00000000fcfd42fa C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00279290 iqm::Iqm::setupIqmFcnPtrs+00071530 [ 52] 0x00000000fcfd41d3 C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00278995 iqm::Iqm::setupIqmFcnPtrs+00071235 [ 53] 0x00000000fcfb9e5a C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00171610 iqm::Iqm::deliver+00003274 [ 54] 0x00000000fcfbad2f C:\Program Files\MATLAB\R2016b\bin\win64\iqm.dll+00175407 iqm::Iqm::deliver+00007071 [ 55] 0x0000000100108173 C:\Program Files\MATLAB\R2016b\bin\win64\libmwservices.dll+01147251 services::system_events::PpeDispatchHook::dispatchOne+00019059 [ 56] 0x000000010010c3a3 C:\Program Files\MATLAB\R2016b\bin\win64\libmwservices.dll+01164195 sysq::addProcessPendingEventsUnitTestHook+00001923 [ 57] 0x000000010010c580 C:\Program Files\MATLAB\R2016b\bin\win64\libmwservices.dll+01164672 sysq::addProcessPendingEventsUnitTestHook+00002400 [ 58] 0x000000010010d685 C:\Program Files\MATLAB\R2016b\bin\win64\libmwservices.dll+01169029 sysq::getCondition+00002917 [ 59] 0x000000010010e4b2 C:\Program Files\MATLAB\R2016b\bin\win64\libmwservices.dll+01172658 svWS_ProcessPendingEvents+00000162 [ 60] 0x0000000007f96b7d C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00224125 mcr::runtime::setInterpreterThreadToCurrent+00026413 [ 61] 0x0000000007f9726a C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00225898 mcr::runtime::setInterpreterThreadToCurrent+00028186 [ 62] 0x0000000007f8e385 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00189317 mcr_process_events+00010293 [ 63] 0x0000000007f902d2 C:\Program Files\MATLAB\R2016b\bin\win64\mcr.dll+00197330 mcr_process_events+00018306 [ 64] 0x0000000007e3bd0e C:\Program Files\MATLAB\R2016b\bin\win64\MVMLocal.dll+00245006 mvm_server::inproc::LocalFactory::terminate+00070846 [ 65] 0x00000000fa942cd9 C:\Program Files\MATLAB\R2016b\bin\win64\mvm.dll+01125593 mvm::detail::initLocalMvmHack+00000521 [ 66] 0x00000000fa9432a5 C:\Program Files\MATLAB\R2016b\bin\win64\mvm.dll+01127077 mvm::detail::SessionImpl::privateSession+00000373 [ 67] 0x00000000fa9434b1 C:\Program Files\MATLAB\R2016b\bin\win64\mvm.dll+01127601 mvm::detail::SessionImpl::privateSession+00000897 [ 68] 0x0000000140006fd5 C:\Program Files\MATLAB\R2016b\bin\win64\MATLAB.exe+00028629 [ 69] 0x0000000140007661 C:\Program Files\MATLAB\R2016b\bin\win64\MATLAB.exe+00030305 [ 70] 0x00007ffbb49213d2 C:\Windows\system32\KERNEL32.DLL+00005074 BaseThreadInitThunk+00000034 [ 71] 0x00007ffbb4a754e4 C:\Windows\SYSTEM32\ntdll.dll+00087268 RtlUserThreadStart+00000052
This error was detected while a MEX-file was running. If the MEX-file is not an official MathWorks function, please examine its source code for errors. Please consult the External Interfaces Guide for information on debugging MEX-files.
If this problem is reproducible, please submit a Service Request via: http://www.mathworks.com/support/contact_us/
A technical support engineer might contact you with further information.
Thank you for your help.
It was difficult to figure out the exact point getting crash because it collapsed without any error message on commend window But I tried to figure out that.
The route is like this;
Net.m line 171 function reshape(self) caffe_('net_reshape', self.hNet.self);
Net.m line 186 function reshape_as_input_data self.reshape();
fast_rcnn_conv_feat_detect.m line 33 caffe_net.reshape_as_input(net_inputs);
scrip_faster_rcnn_demo.m line 59 [boxes, scores] fast_rcnn_conv_feat_detect(proposal_detection_model.conf_detection, fast_rcnn_net, im, rpn_net.blobs(proposal_detection_model.last_shared_output_blob_name), aboxes(:, 1:4), opts.after_nms_topN);
I'm not expert on this, so I got trouble on it. I don't know why pre-trained model goes on well but self-trained model doesn't go on well even though datasets are not different.
Additionally, if there is any kind an archive of pre-trained model for RCNN, I'll glad to let me know. My purpose is just obtaining the tuned model for my research.
@INGslh There's difference in the script(I don't know how to call it exactly) between those of pre-trained model and self-trained model in the output folder. I just tried copying some part of script from pre-trained one to self-trained one, then it works without any error.
I did this, but the test results are poor, what method do you have to solve this problem?
@INGslh The files are 'detection_test.prototxt' and 'proposal_test.prototxt'. But I'm not sure it could be the solution for yours. Because you said that 'yours results were poor'. You already got the results. What I mentioned is the solution for crash.