dutran
commented
6 years ago @nicebaby You may want to extract your videos UCF101/UCF-101/JumpingJack/v_JumpingJack_g25_c05.avi into frames, e.g. UCF101/UCF-101/JumpingJack/v_JumpingJack_g25_c05/%06d.jpg before you do fine-tuning
when I run the ucf101_finetuning.sh(examples/c3d_finetuning) ,the following question is:
I1224 17:09:40.411164 7406 video_data_layer.cpp:317] A total of 107258 video chunks. E1224 17:09:40.411537 7406 image_io.cpp:111] Cannot open UCF101/UCF-101/JumpingJack/v_JumpingJack_g25_c05.avi/ F1224 17:09:40.411547 7406 video_data_layer.cpp:341] Check failed: ReadVideoToVolumeDatum(filelist[id].c_str(), start_frmlist[id], labellist[id], new_length, new_height, new_width, sampling_rate, &datum) Check failure stack trace: @ 0x7f0fbb09a5cd google::LogMessage::Fail() @ 0x7f0fbb09c433 google::LogMessage::SendToLog() @ 0x7f0fbb09a15b google::LogMessage::Flush() @ 0x7f0fbb09ce1e google::LogMessageFatal::~LogMessageFatal() @ 0x4e32b6 caffe::VideoDataLayer<>::SetUp() @ 0x43efb7 caffe::Net<>::Init() @ 0x440475 caffe::Net<>::Net() @ 0x44455f caffe::Solver<>::Init() @ 0x44736b caffe::Solver<>::Solver() @ 0x40c83d main @ 0x7f0fb326b830 __libc_start_main @ 0x40ea89 _start @ (nil) (unknown) ./ucf101_finetuning.sh: 行 1: 7406 已放弃 (核心已转储) GLOG_logtostderr=1 ../../build/tools/finetune_net.bin c3d_ucf101_finetuning_solver.prototxt ../c3d_feature_extraction/conv3d_deepnetA_sport1m_iter_1900000 Tried many ways, but still no solution.I would appreciated for some advice,very thanks!