当前位置：网站首页>Installing C3d v1.0

Installing C3d v1.0

2022-07-21 16:36:00 【MallocLu】

Preface

Main reference course

Official website

official User Guide

official github

personal caffe Installation tutorial

Files or compressed packages that may be used ：

link ：https://pan.baidu.com/s/1vIRxSJHRoBnrnsjczISOPg?pwd=7asr
Extraction code ：7asr
Copy this content after opening baidu network disk mobile phone App, It is more convenient to operate

All the following commands are in root Execute under authority , If the prompt authority is insufficient , Then add sudo

Main reference tutorial if the website fails , You can see it in the net disk “ Main reference tutorial web page backup ”

This article introduces C3D v1.0, If you want to use C3D v1.1, Then it and caffe The installation process is exactly the same , follow “ personal caffe Installation tutorial ” Go over , And then install it C3D v1.1 that will do

install CUDA&CUDNN

install CUDA 10.2 And the corresponding CUDNN（ Reference resources “ Main reference course ”）, In order to simplify the operation, the author uses “ Moment pool cloud platform ”：

It's already configured CUDA and cuDNN, You don't have to configure it yourself . Open one 1 element / When the environment is enough

Installation dependency

apt-get update
apt-get install -y libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
apt-get install --no-install-recommends libboost-all-dev
apt-get install -y libopenblas-dev liblapack-dev libatlas-base-dev libgflags-dev libgoogle-glog-dev liblmdb-dev git cmake build-essential

#  Open the environment profile 
vim ~/.bashrc

#  Add the following two lines at the end of the open file 
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH

#  Make the environment configuration file effective 
source ~/.bashrc

install Opencv

take opencv-3.4.10.zip Copied to the /root Under the table of contents
unzip opencv-3.4.10.zip
Delete opencv-3.4.10.zip（ To save space ）

Execute the following command

#  Enter folder 
cd opencv-3.4.10
#  establish build Folder 
mkdir build
#  Get into build Folder 
cd build
#  compile 
cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local ..
make
#  install 
make install
#  Check whether the installation is successful   Success will output ：3.4.10
pkg-config --modversion opencv

Default Python2

alias python=python2
#  Output python Version is 2.? It's success 
python -V

install boost

take boost_1_72_0.7z Copied to the /root Under the table of contents
apt-get install p7zip-full
7z x boost_1_72_0.7z
Delete boost_1_72_0.7z（ To save space ）

Execute the following command

#  Enter folder 
cd boost_1_72_0.7z

./bootstrap.sh --with-libraries=all 
./b2
./b2 install

install C3D

take C3D-master.zip Copied to the /root Under the table of contents
unzip C3D-master.zip
Delete C3D-master.zip（ To save space ）

Execute the following command

#  Enter folder 
cd C3D-master/C3D-v1.0/python
#  Install dependency packages 
for req in $(cat requirements.txt); do pip install $req; done
#  Return to the previous directory 
cd ..
#  Copy file 
cp Makefile.config.example Makefile.config

Modify file

1） modify Makefile.config

CUDA_ARCH := -gencode arch=compute_20,code=sm_20 \
		-gencode arch=compute_20,code=sm_21 \
		-gencode arch=compute_30,code=sm_30 \
		-gencode arch=compute_35,code=sm_35 \
		-gencode=arch=compute_50,code=sm_50 \
		#-gencode=arch=compute_50,code=compute_50 
 It is amended as follows 
CUDA_ARCH := #-gencode arch=compute_20,code=sm_20 \
		#-gencode arch=compute_20,code=sm_21 \
		#-gencode arch=compute_30,code=sm_30 \
		#-gencode arch=compute_35,code=sm_35 \
		-gencode=arch=compute_50,code=sm_50 \
		-gencode=arch=compute_50,code=compute_50
		
# OPENCV_VERSION := 3
 It is amended as follows 
OPENCV_VERSION := 3

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib 
 It is amended as follows 
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial

2） modify Makefile

NVCCFLAGS +=-ccbin=$(CXX) -Xcompiler-fPIC $(COMMON_FLAGS)
 It is amended as follows 
NVCCFLAGS += -D_FORCE_INLINES -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)

LIBRARIES := cudart cublas curand \
	pthread \
	glog protobuf leveldb snappy \
	boost_system \
	hdf5_hl hdf5 \
	opencv_core opencv_highgui opencv_imgproc
 It is amended as follows 
LIBRARIES := cudart cublas curand \
	pthread \
	glog protobuf leveldb snappy \
	boost_system \
	hdf5_serial_hl hdf5_serial \
	opencv_core opencv_highgui opencv_imgproc

3） modify /src/caffe/test/test_gradient_check_util.hpp

Dtype scale = std::max(
std::max(fabs(computed_gradient), fabs(estimated_gradient)), 1.);
 It is amended as follows 
Dtype scale = std::max(
std::max(fabs(computed_gradient), fabs(estimated_gradient)),
Dtype(1.));

compile

make all
#  There may be an error ：recipe for target 'build/examples/mnist/convert_mnist_data.o' failed
#  Never mind 

make test

make runtest
#  There may be an error , Never mind 
# Makefile:285: recipe for target 'runtest' failed
# make: *** [runtest] Error 127

C3D test

Will file conv3d_deepnetA_sport1m_iter_1900000 Copied to the C3D-master/C3D-v1.0/examples/c3d_feature_extraction/
sh c3d_sport1m_feature_extraction_frm.sh
There is an error ：…/…/build/tools/extract_image_features.bin: error while loading shared libraries: libopencv_core.so.3.4: cannot open shared object file: No such file or directory
solve ： stay /etc/ld.so.conf.d/ Create OpenCV.conf, The content is /usr/local/lib, And then execute the command sudo ldconfig
sh c3d_sport1m_feature_extraction_frm.sh
Successful implementation ,E0707 14:27:41.620281 36381 extract_image_features.cpp:121] Successfully extracted 16 features!