search

ethanhe42 / channel-pruning

Channel Pruning for Accelerating Very Deep Neural Networks (ICCV'17)
https://arxiv.org/abs/1707.06168
MIT License
1.07k stars 310 forks source link

Check failed: indices_to_forward_.size() == new_tops_num (64 vs. 35)? #98

Closed Eric-mingjie closed 5 years ago

Eric-mingjie commented 6 years ago

What steps reproduce the bug?

Download the Resnet-50 channel pruning model. Then use the caffe suggested here to train the model resnet-50 with the model prototxt.

If the bug is a crash, provide the backtrace.

F0712 02:17:08.466684  1261 filter_layer.cu:39] Check failed: indices_to_forward_.size() == new_tops_num (64 vs. 35) 
*** Check failure stack trace: ***
    @     0x7f8c7d9b75cd  google::LogMessage::Fail()
    @     0x7f8c7d9b9433  google::LogMessage::SendToLog()
F0712 02:17:08.476307  1262 filter_layer.cu:39] Check failed: indices_to_forward_.size() == new_tops_num (64 vs. 35) 
*** Check failure stack trace: ***
    @     0x7f8c7d9b715b  google::LogMessage::Flush()
    @     0x7f8c7d9b75cd  google::LogMessage::Fail()
    @     0x7f8c7d9b9e1e  google::LogMessageFatal::~LogMessageFatal()
    @     0x7f8c7d9b9433  google::LogMessage::SendToLog()
F0712 02:17:08.476307  1262 filter_layer.cu:39] Check failed: indices_to_forward_.size() == new_tops_num (64 vs. 35) F0712 02:17:08.483189  1260 filter_layer.cu:39] Check failed: indices_to_forward_.size() == new_tops_num (64 vs. 35) 
*** Check failure stack trace: ***
    @     0x7f8c7d9b715b  google::LogMessage::Flush()
    @     0x7f8c7e1b96a6  caffe::FilterLayer<>::Forward_gpu()
    @     0x7f8c7d9b9e1e  google::LogMessageFatal::~LogMessageFatal()
    @     0x7f8c7d9b75cd  google::LogMessage::Fail()
    @     0x7f8c7e1b96a6  caffe::FilterLayer<>::Forward_gpu()
    @     0x7f8c7e1978b1  caffe::Net<>::ForwardFromTo()
    @     0x7f8c7e1978b1  caffe::Net<>::ForwardFromTo()
    @     0x7f8c7d9b9433  google::LogMessage::SendToLog()
    @     0x7f8c7e1979b7  caffe::Net<>::Forward()
    @     0x7f8c7e1979b7  caffe::Net<>::Forward()
    @     0x7f8c7e18e8d8  caffe::Solver<>::Step()
    @     0x7f8c7d9b715b  google::LogMessage::Flush()
    @     0x7f8c7e1254be  caffe::Worker<>::InternalThreadEntry()
    @     0x7f8c7e18e8d8  caffe::Solver<>::Step()
    @     0x7f8c7e110045  caffe::InternalThread::entry()
    @     0x7f8c7d9b9e1e  google::LogMessageFatal::~LogMessageFatal()
    @     0x7f8c7e110f6e  boost::detail::thread_data<>::run()
    @     0x7f8c7e1254be  caffe::Worker<>::InternalThreadEntry()
    @     0x7f8c7345c5d5  (unknown)
    @     0x7f8c7e1b96a6  caffe::FilterLayer<>::Forward_gpu()
    @     0x7f8c5f7336ba  start_thread
    @     0x7f8c7e110045  caffe::InternalThread::entry()
    @     0x7f8c7c13c3dd  clone
    @     0x7f8c7e1978b1  caffe::Net<>::ForwardFromTo()
    @              (nil)  (unknown)
Aborted (core dumped)
ethanhe42 commented 5 years ago

Please check again. I can't reproduce this error. Your .prototxt didn't match .caffemodel.