Kezunlin's Blog

Live and Learn

0%

Install and Configure Caffe on ubuntu 16.04

Posted on Edited on In

Series

Guide

requirements:

  • NVIDIA driver 396.54
  • CUDA 8.0 + cudnn 6.0.21
  • CUDA 9.2 +cudnn 7.1.4
  • opencv 3.1.0 —>3.3.0
  • python 2.7 + numpy 1.15.1
  • python 3.5.2 + numpy 1.16.2
  • protobuf 3.6.1 （static）
  • caffe latest

默认的protobuf,2.6.1测试通过。
此处，使用最新的3.6.1 也可以，编译caffe需要加上-std=c++11

install CUDA + cudnn

see install and configure cuda 9.2 with cudnn 7.1 on ubuntu 16.04
tips: we need to recompile caffe with cudnn 7.1

before we compile caffe, move caffe/python/caffe/selective_search_ijcv_with_python folder outside caffe source folder, otherwise error occurs.

recompile caffe with cudnn

install protobuf

see Part 1: compile protobuf-cpp on ubuntu 16.04

1
2
3
4
5
which protoc 
/usr/local/bin/protoc

protoc --version
libprotoc 3.6.1

caffe使用static的libprotoc 3.6.1

install opencv

see compile opencv on ubuntu 16.04

1
2
3
4
5
which opencv_version
/usr/local/bin/opencv_version

opencv_version 
3.3.0

python

1
2
python --version
Python 2.7.12

check numpy version

1
2
3
4
5
6
7
8
import numpy
numpy.__version__
'1.15.1'

import numpy
import inspect
inspect.getfile(numpy)
'/usr/local/lib/python2.7/dist-packages/numpy/__init__.pyc'

compile caffe

clone repo

1
2
git clone https://github.com/BVLC/caffe.git 
cd caffe

update repo

update at 20180822.

if you change your local Makefile and git pull origin master merge conflict, solution

1
2
git checkout HEAD Makefile
git pull origin master

configure

1
mkdir build && cd build && cmake-gui ..

cmake-gui options

USE_CUDNN ON
USE_OPENCV ON
Build_python ON
Build_python_layer ON

BLAS atlas
CMAKE_CXX_FLGAS -std=c++11

CMAKE_INSTALL_PREFIX /home/kezunlin/program/caffe/build/install

使用-std=c++11

configure output 

Dependencies:
  BLAS              :   Yes (Atlas)
  Boost             :   Yes (ver. 1.66)
  glog              :   Yes
  gflags            :   Yes
  protobuf          :   Yes (ver. 3.6.1)
  lmdb              :   Yes (ver. 0.9.17)
  LevelDB           :   Yes (ver. 1.18)
  Snappy            :   Yes (ver. 1.1.3)
  OpenCV            :   Yes (ver. 3.1.0)
  CUDA              :   Yes (ver. 9.2)

NVIDIA CUDA:
  Target GPU(s)     :   Auto
  GPU arch(s)       :   sm_61
  cuDNN             :   Yes (ver. 7.1.4)

Python:
  Interpreter       :   /usr/bin/python2.7 (ver. 2.7.12)
  Libraries         :   /usr/lib/x86_64-linux-gnu/libpython2.7.so (ver 2.7.12)
  NumPy             :   /usr/lib/python2.7/dist-packages/numpy/core/include (ver 1.51.1)

Documentaion:
  Doxygen           :   /usr/bin/doxygen (1.8.11)
  config_file       :   /home/kezunlin/program/caffe/.Doxyfile

Install:
  Install path      :   /home/kezunlin/program/caffe/build/install

Configuring done

we can also use python3.5 and numpy 1.16.2

Python:
  Interpreter       :   /usr/bin/python3 (ver. 3.5.2)
  Libraries         :   /usr/lib/x86_64-linux-gnu/libpython3.5m.so (ver 3.5.2)
  NumPy             :   /home/kezunlin/.local/lib/python3.5/site-packages/numpy/core/include (ver 1.16.2)

use -std=c++11, otherwise errors occur

make -j8
[  1%] Running C++/Python protocol buffer compiler on /home/kezunlin/program/caffe/src/caffe/proto/caffe.proto
Scanning dependencies of target caffeproto
[  1%] Building CXX object src/caffe/CMakeFiles/caffeproto.dir/__/__/include/caffe/proto/caffe.pb.cc.o
In file included from /usr/include/c++/5/mutex:35:0,
                 from /usr/local/include/google/protobuf/stubs/mutex.h:33,
                 from /usr/local/include/google/protobuf/stubs/common.h:52,
                 from /home/kezunlin/program/caffe/build/include/caffe/proto/caffe.pb.h:9,
                 from /home/kezunlin/program/caffe/build/include/caffe/proto/caffe.pb.cc:4:
/usr/include/c++/5/bits/c++0x_warning.h:32:2: error: #error This file requires compiler and library support for the ISO C++ 2011 standard. This support must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
 #error This file requires compiler and library support \

fix gcc error

edit /usr/local/cuda/include/host_config.h

将其中的第115行注释掉：

1
2
3
#error-- unsupported GNU version! gcc versions later than 4.9 are not supported!
======>
//#error-- unsupported GNU version! gcc versions later than 4.9 are not supported!

fix gflags error

  • caffe/include/caffe/common.hpp
  • caffe/examples/mnist/convert_mnist_data.cpp

Comment out the ifndef

1
2
3
// #ifndef GFLAGS_GFLAGS_H_
namespace gflags = google;
// #endif  // GFLAGS_GFLAGS_H_

compile

1
2
3
make clean
make -j8 
make pycaffe

output

[  1%] Running C++/Python protocol buffer compiler on /home/kezunlin/program/caffe/src/caffe/proto/caffe.proto
Scanning dependencies of target caffeproto
[  1%] Building CXX object src/caffe/CMakeFiles/caffeproto.dir/__/__/include/caffe/proto/caffe.pb.cc.o
[  1%] Linking CXX static library ../../lib/libcaffeproto.a
[  1%] Built target caffeproto

libcaffeproto.a　static library

install

1
2
3
4
make install

ls build/install
bin  include  lib  python  share

install to build/install folder

1
2
ls build/install/lib
libcaffeproto.a  libcaffe.so  libcaffe.so.1.0.0

advanced

  • INTERFACE_INCLUDE_DIRECTORIES
  • INTERFACE_LINK_LIBRARIES

Target “caffe” has an INTERFACE_LINK_LIBRARIES property which differs from its LINK_INTERFACE_LIBRARIES properties.

Play with Caffe

python caffe

fix python caffe

fix ipython 6.1 version conflict

vim caffe/python/requirements.txt

1
2
3
ipython>=3.0.0
====>
ipython==5.4.1

reinstall ipython

1
2
3
4
5
pip install -r requirements.txt

cd caffe/python
python
>>>import caffe

python draw net

1
2
3
4
5
6
7
8
9
10
11
sudo apt-get install graphviz
sudo pip install theano=0.9

# for theano d3viz
sudo pip install pydot==1.1.0
sudo pip install pydot-ng


# other usefull tools
sudo pip install jupyter
sudo pip install seaborn

we need to install graphviz, otherwise we get ERROR:”dot” not found in path

draw net 

1
2
3
4
cd $CAFFE_HOME
./python/draw_net.py ./examples/mnist/lenet.prototxt ./examples/mnist/lenet.png

eog ./examples/mnist/lenet.png

lenet

cpp caffe

train net

1
2
3
4
5
6
cd caffe
./examples/mnist/create_mnist.sh
./examples/mnist/train_lenet.sh 

cat ./examples/mnist/train_lenet.sh 
./build/tools/caffe train --solver=examples/mnist/lenet_solver.prototxt $@

output results

I0912 15:57:28.812655 14094 solver.cpp:327] Iteration 10000, loss = 0.00272129
I0912 15:57:28.812675 14094 solver.cpp:347] Iteration 10000, Testing net (#0)
I0912 15:57:28.891481 14100 data_layer.cpp:73] Restarting data prefetching from start.
I0912 15:57:28.893678 14094 solver.cpp:414]     Test net output #0: accuracy = 0.9904
I0912 15:57:28.893707 14094 solver.cpp:414]     Test net output #1: loss = 0.0276084 (* 1 = 0.0276084 loss)
I0912 15:57:28.893714 14094 solver.cpp:332] Optimization Done.
I0912 15:57:28.893719 14094 caffe.cpp:250] Optimization Done.

tips, for caffe, errors because no imdb data.

I0417 13:28:17.764714 35030 layer_factory.hpp:77] Creating layer mnist
F0417 13:28:17.765067 35030 db_lmdb.hpp:15] Check failed: mdb_status == 0 (2 vs. 0) No such file or directory
---------------------

upgrade net

1
2
./tools/upgrade_net_proto_text  old.prototxt new.prototxt
./tools/upgrade_net_proto_binary  old.caffemodel new.caffemodel

caffe time

yolov3

1
2
3
4
5
./build/tools/caffe time --model='det/yolov3/yolov3.prototxt' --iterations=100 --gpu=0

I0313 10:15:41.888208 12527 caffe.cpp:408] Average Forward pass: 49.7012 ms.
I0313 10:15:41.888213 12527 caffe.cpp:410] Average Backward pass: 84.946 ms.
I0313 10:15:41.888248 12527 caffe.cpp:412] Average Forward-Backward: 134.85 ms.

yolov3 5 class

1
2
3
4
5
./build/tools/caffe time --model='det/autotrain/yolo3-5c.prototxt' --iterations=100 --gpu=0

I0313 10:19:27.283625 12894 caffe.cpp:408] Average Forward pass: 38.4823 ms.
I0313 10:19:27.283630 12894 caffe.cpp:410] Average Backward pass: 74.1656 ms.
I0313 10:19:27.283638 12894 caffe.cpp:412] Average Forward-Backward: 112.732 ms.

Example

Caffe Classifier

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
#include <caffe/caffe.hpp>
#ifdef USE_OPENCV
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#endif  // USE_OPENCV
#include <algorithm>
#include <iosfwd>
#include <memory>
#include <string>
#include <utility>
#include <vector>

#ifdef USE_OPENCV
using namespace caffe;  // NOLINT(build/namespaces)
using std::string;

/* Pair (label, confidence) representing a prediction. */
typedef std::pair<string, float> Prediction;

class Classifier {
 public:
  Classifier(const string& model_file,
             const string& trained_file,
             const string& mean_file,
             const string& label_file);

  std::vector<Prediction> Classify(const cv::Mat& img, int N = 5);

 private:
  void SetMean(const string& mean_file);

  std::vector<float> Predict(const cv::Mat& img);

  void WrapInputLayer(std::vector<cv::Mat>* input_channels);

  void Preprocess(const cv::Mat& img,
                  std::vector<cv::Mat>* input_channels);

 private:
  shared_ptr<Net<float> > net_;
  cv::Size input_geometry_;
  int num_channels_;
  cv::Mat mean_;
  std::vector<string> labels_;
};

Classifier::Classifier(const string& model_file,
                       const string& trained_file,
                       const string& mean_file,
                       const string& label_file) {
#ifdef CPU_ONLY
  Caffe::set_mode(Caffe::CPU);
#else
  Caffe::set_mode(Caffe::GPU);
#endif

  /* Load the network. */
  net_.reset(new Net<float>(model_file, TEST));
  net_->CopyTrainedLayersFrom(trained_file);

  CHECK_EQ(net_->num_inputs(), 1) << "Network should have exactly one input.";
  CHECK_EQ(net_->num_outputs(), 1) << "Network should have exactly one output.";

  Blob<float>* input_layer = net_->input_blobs()[0];
  num_channels_ = input_layer->channels();
  CHECK(num_channels_ == 3 || num_channels_ == 1)
    << "Input layer should have 1 or 3 channels.";
  input_geometry_ = cv::Size(input_layer->width(), input_layer->height());

  /* Load the binaryproto mean file. */
  SetMean(mean_file);

  /* Load labels. */
  std::ifstream labels(label_file.c_str());
  CHECK(labels) << "Unable to open labels file " << label_file;
  string line;
  while (std::getline(labels, line))
    labels_.push_back(string(line));

  Blob<float>* output_layer = net_->output_blobs()[0];
  CHECK_EQ(labels_.size(), output_layer->channels())
    << "Number of labels is different from the output layer dimension.";
}

static bool PairCompare(const std::pair<float, int>& lhs,
                        const std::pair<float, int>& rhs) {
  return lhs.first > rhs.first;
}

/* Return the indices of the top N values of vector v. */
static std::vector<int> Argmax(const std::vector<float>& v, int N) {
  std::vector<std::pair<float, int> > pairs;
  for (size_t i = 0; i < v.size(); ++i)
    pairs.push_back(std::make_pair(v[i], i));
  std::partial_sort(pairs.begin(), pairs.begin() + N, pairs.end(), PairCompare);

  std::vector<int> result;
  for (int i = 0; i < N; ++i)
    result.push_back(pairs[i].second);
  return result;
}

/* Return the top N predictions. */
std::vector<Prediction> Classifier::Classify(const cv::Mat& img, int N) {
  std::vector<float> output = Predict(img);

  N = std::min<int>(labels_.size(), N);
  std::vector<int> maxN = Argmax(output, N);
  std::vector<Prediction> predictions;
  for (int i = 0; i < N; ++i) {
    int idx = maxN[i];
    predictions.push_back(std::make_pair(labels_[idx], output[idx]));
  }

  return predictions;
}

/* Load the mean file in binaryproto format. */
void Classifier::SetMean(const string& mean_file) {
  BlobProto blob_proto;
  ReadProtoFromBinaryFileOrDie(mean_file.c_str(), &blob_proto);

  /* Convert from BlobProto to Blob<float> */
  Blob<float> mean_blob;
  mean_blob.FromProto(blob_proto);
  CHECK_EQ(mean_blob.channels(), num_channels_)
    << "Number of channels of mean file doesn't match input layer.";

  /* The format of the mean file is planar 32-bit float BGR or grayscale. */
  std::vector<cv::Mat> channels;
  float* data = mean_blob.mutable_cpu_data();
  for (int i = 0; i < num_channels_; ++i) {
    /* Extract an individual channel. */
    cv::Mat channel(mean_blob.height(), mean_blob.width(), CV_32FC1, data);
    channels.push_back(channel);
    data += mean_blob.height() * mean_blob.width();
  }

  /* Merge the separate channels into a single image. */
  cv::Mat mean;
  cv::merge(channels, mean);

  /* Compute the global mean pixel value and create a mean image
   * filled with this value. */
  cv::Scalar channel_mean = cv::mean(mean);
  mean_ = cv::Mat(input_geometry_, mean.type(), channel_mean);
}

std::vector<float> Classifier::Predict(const cv::Mat& img) {
  Blob<float>* input_layer = net_->input_blobs()[0];
  input_layer->Reshape(1, num_channels_,
                       input_geometry_.height, input_geometry_.width);
  /* Forward dimension change to all layers. */
  net_->Reshape();

  std::vector<cv::Mat> input_channels;
  WrapInputLayer(&input_channels);

  Preprocess(img, &input_channels);

  net_->Forward();

  /* Copy the output layer to a std::vector */
  Blob<float>* output_layer = net_->output_blobs()[0];
  const float* begin = output_layer->cpu_data();
  const float* end = begin + output_layer->channels();
  return std::vector<float>(begin, end);
}

/* Wrap the input layer of the network in separate cv::Mat objects
 * (one per channel). This way we save one memcpy operation and we
 * don't need to rely on cudaMemcpy2D. The last preprocessing
 * operation will write the separate channels directly to the input
 * layer. */
void Classifier::WrapInputLayer(std::vector<cv::Mat>* input_channels) {
  Blob<float>* input_layer = net_->input_blobs()[0];

  int width = input_layer->width();
  int height = input_layer->height();
  float* input_data = input_layer->mutable_cpu_data();
  for (int i = 0; i < input_layer->channels(); ++i) {
    cv::Mat channel(height, width, CV_32FC1, input_data);
    input_channels->push_back(channel);
    input_data += width * height;
  }
}

void Classifier::Preprocess(const cv::Mat& img,
                            std::vector<cv::Mat>* input_channels) {
  /* Convert the input image to the input image format of the network. */
  cv::Mat sample;
  if (img.channels() == 3 && num_channels_ == 1)
    cv::cvtColor(img, sample, cv::COLOR_BGR2GRAY);
  else if (img.channels() == 4 && num_channels_ == 1)
    cv::cvtColor(img, sample, cv::COLOR_BGRA2GRAY);
  else if (img.channels() == 4 && num_channels_ == 3)
    cv::cvtColor(img, sample, cv::COLOR_BGRA2BGR);
  else if (img.channels() == 1 && num_channels_ == 3)
    cv::cvtColor(img, sample, cv::COLOR_GRAY2BGR);
  else
    sample = img;

  cv::Mat sample_resized;
  if (sample.size() != input_geometry_)
    cv::resize(sample, sample_resized, input_geometry_);
  else
    sample_resized = sample;

  cv::Mat sample_float;
  if (num_channels_ == 3)
    sample_resized.convertTo(sample_float, CV_32FC3);
  else
    sample_resized.convertTo(sample_float, CV_32FC1);

  cv::Mat sample_normalized;
  cv::subtract(sample_float, mean_, sample_normalized);

  /* This operation will write the separate BGR planes directly to the
   * input layer of the network because it is wrapped by the cv::Mat
   * objects in input_channels. */
  cv::split(sample_normalized, *input_channels);

  CHECK(reinterpret_cast<float*>(input_channels->at(0).data)
        == net_->input_blobs()[0]->cpu_data())
    << "Input channels are not wrapping the input layer of the network.";
}

int main(int argc, char** argv) {
  if (argc != 6) {
    std::cerr << "Usage: " << argv[0]
              << " deploy.prototxt network.caffemodel"
              << " mean.binaryproto labels.txt img.jpg" << std::endl;
    return 1;
  }

  ::google::InitGoogleLogging(argv[0]);

  string model_file   = argv[1];
  string trained_file = argv[2];
  string mean_file    = argv[3];
  string label_file   = argv[4];
  Classifier classifier(model_file, trained_file, mean_file, label_file);

  string file = argv[5];

  std::cout << "---------- Prediction for "
            << file << " ----------" << std::endl;

  cv::Mat img = cv::imread(file, -1);
  CHECK(!img.empty()) << "Unable to decode image " << file;
  std::vector<Prediction> predictions = classifier.Classify(img);

  /* Print the top N predictions. */
  for (size_t i = 0; i < predictions.size(); ++i) {
    Prediction p = predictions[i];
    std::cout << std::fixed << std::setprecision(4) << p.second << " - \""
              << p.first << "\"" << std::endl;
  }
}
#else
int main(int argc, char** argv) {
  LOG(FATAL) << "This example requires OpenCV; compile with USE_OPENCV.";
}
#endif  // USE_OPENCV

CMakeLists.txt

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
find_package(OpenCV REQUIRED)


set(Caffe_DIR "/home/kezunlin/program/caffe/build/install/share/Caffe")  # caffe caffe

# for CaffeConfig.cmake/ caffe-config.cmake
find_package(Caffe)
# offical caffe  : There is no Caffe_INCLUDE_DIRS and Caffe_DEFINITIONS
# refinedet caffe: OK.

add_definitions(${Caffe_DEFINITIONS})

MESSAGE( [Main] " Caffe_INCLUDE_DIRS = ${Caffe_INCLUDE_DIRS}")
MESSAGE( [Main] " Caffe_DEFINITIONS = ${Caffe_DEFINITIONS}")
MESSAGE( [Main] " Caffe_LIBRARIES = ${Caffe_LIBRARIES}") # caffe
MESSAGE( [Main] " Caffe_CPU_ONLY = ${Caffe_CPU_ONLY}")
MESSAGE( [Main] " Caffe_HAVE_CUDA = ${Caffe_HAVE_CUDA}")
MESSAGE( [Main] " Caffe_HAVE_CUDNN = ${Caffe_HAVE_CUDNN}")


include_directories(${Caffe_INCLUDE_DIRS})

target_link_libraries(demo 
    ${OpenCV_LIBS}
    ${Caffe_LIBRARIES}
)

run

1
./demo

if error occurs:

libcaffe.so.1.0.0 => not found

edit .bashrc

1
2
# for caffe
export LD_LIBRARY_PATH=/home/kezunlin/program/caffe/build/install/lib:$LD_LIBRARY_PATH

Reference

History

  • 20180807: created.
  • 20180822: update cmake-gui for caffe