Ubuntu18.04
gcc-5.5
g++-5.5
cudn9.0
cudnn7.4
opencv3.4
matlab2017a
protobuf2.6.1
caffe
caffe-fast-rcnn
faster-rcnn-cpu
Just compile it using the following command:
sudo make all -j16
sudo make test -j16
sudo make pycaffe
sudo make matcaffe -j16
sudo make mattest -j16
If you don't want to download caffe,You can copy files to your corresponding path from 'caffe_to_fast_rcnn(cpu)' to 'caffe-fast-rcnn'.
caffe-fast-rcnn-compile.sh
root@gmt-ThinkPad-T470p:/home/gmt/caffe-fast-rcnn# sudo make all -j16
PROTOC src/caffe/proto/caffe.proto
NVCC src/caffe/layers/relu_layer.cu
...................................
CXX/LD -o .build_release/examples/mnist/convert_mnist_data.bin
root@gmt-ThinkPad-T470p:/home/gmt/caffe-fast-rcnn# sudo make test -j16
....................................
CXX/LD -o .build_release/test/test_all.testbin src/caffe/test/test_caffe_main.cpp
root@gmt-ThinkPad-T470p:/home/gmt/caffe-fast-rcnn# sudo make pycaffe
....................................
PROTOC (python) src/caffe/proto/caffe.proto
root@gmt-ThinkPad-T470p:/home/gmt/caffe-fast-rcnn# sudo make matcaffe -j16
MEX matlab/+caffe/private/caffe_.cpp
MEX 已成功完成。
root@gmt-ThinkPad-T470p:/home/gmt/caffe-fast-rcnn# sudo make mattest -j16
Cleared 2 solvers and 4 stand-alone nets
ans =
1×7 TestResult 数组 - 属性:
Name
Passed
Failed
Incomplete
Duration
Details
总计:
7 Passed, 0 Failed, 0 Incomplete.
0.50392 秒测试时间。
1.Ubuntu18.04 install
2.gcc-5/g++-5 and gcc-7/g++-7 compatible install
3.nvidia390 drivers,cuda9.0,cudnn7.4
4.ubuntu+cuda9.0+cudnn7+opencv+matlab+caffe
5.opencv install and opencv uninstall
6.matlab2017a
7.protobuf error same as #5078 tsing-cv and protobuf install by reading the README file in protobuf-2.6.0
8.caffe
<---------------------------------------wait for update------------------------------------------->