Open hrexha opened 3 years ago
Hi, I tried to finetune the MobileNetSSD on a custom dataset but when it comes to training the script fails with the following message:
net.cpp:850] Check failed: target_blobs.size() >= source_layer.blobs_size() && target_blobs.size() <= source_layer.blobs_size() + 1 Incompatible number of blobs for layer conv0/bn Check failure stack trace: @ 0x7f8d4e5754dd google::LogMessage::Fail() @ 0x7f8d4e57d071 google::LogMessage::SendToLog() @ 0x7f8d4e574ecd google::LogMessage::Flush() @ 0x7f8d4e57676a google::LogMessageFatal::~LogMessageFatal() @ 0x7f8d4d1ac7aa caffe::Net<>::CopyTrainedLayersFrom() @ 0x7f8d4d1b3692 caffe::Net<>::CopyTrainedLayersFromBinaryProto() @ 0x7f8d4d1b371d caffe::Net<>::CopyTrainedLayersFrom() @ 0x55c235b9ee27 CopyLayers() @ 0x55c235ba0688 train() @ 0x55c235b99a49 main @ 0x7f8d4b990b97 __libc_start_main @ 0x55c235b9a689 (unknown) Aborted (core dumped) any idea why is that? Thank you.
Hi, I tried to finetune the MobileNetSSD on a custom dataset but when it comes to training the script fails with the following message:
net.cpp:850] Check failed: target_blobs.size() >= source_layer.blobs_size() && target_blobs.size() <= source_layer.blobs_size() + 1 Incompatible number of blobs for layer conv0/bn Check failure stack trace: @ 0x7f8d4e5754dd google::LogMessage::Fail() @ 0x7f8d4e57d071 google::LogMessage::SendToLog() @ 0x7f8d4e574ecd google::LogMessage::Flush() @ 0x7f8d4e57676a google::LogMessageFatal::~LogMessageFatal() @ 0x7f8d4d1ac7aa caffe::Net<>::CopyTrainedLayersFrom() @ 0x7f8d4d1b3692 caffe::Net<>::CopyTrainedLayersFromBinaryProto() @ 0x7f8d4d1b371d caffe::Net<>::CopyTrainedLayersFrom() @ 0x55c235b9ee27 CopyLayers() @ 0x55c235ba0688 train() @ 0x55c235b99a49 main @ 0x7f8d4b990b97 __libc_start_main @ 0x55c235b9a689 (unknown) Aborted (core dumped) any idea why is that? Thank you.