Installing Caffe on Ubuntu 18.04 with CUDA and CuDNN

Caffe DL Framework

Hi, guys in this tutorial I will go through the steps on installing Caffe on your Linux machine running Ubuntu with support for both CUDA and CuDNN. The Deep Learning Framework Caffe was originally developed by Yangqing Jia at the Vision and Learning Center of the University of California at Berkeley. As Caffe's official site puts it, Caffe should be chosen because of

• Expressive architecture encourages application and innovation. Models and optimization are defined by configuration without hard-coding. Switch between CPU and GPU by setting a single flag to train on a GPU machine then deploy to commodity clusters or mobile devices.
• Extensible code fosters active development. In Caffe’s first year, it has been forked by over 1,000 developers and had many significant changes contributed back. Thanks to these contributors the framework tracks the state-of-the-art in both code and models.
• Speed makes Caffe perfect for research experiments and industry deployment. Caffe can process over 60M images per day with a single NVIDIA K40 GPU*. That’s 1 ms/image for inference and 4 ms/image for learning and more recent library versions and hardware are faster still. We believe that Caffe is among the fastest convnet implementations available.
• Community: Caffe already powers academic research projects, startup prototypes, and even large-scale industrial applications in vision, speech, and multimedia. Join our community of brewers on the caffe-users group and Github.

As you can see Caffe can be a powerful tool when developing convnet models as it’s extremely fast. So let’s start with the installation process.

In this post, we will be building Caffe from the source. I am assuming that you have already installed CUDA 10.0 and CuDNN version 7.6. If you haven’t then please refer to this article. The installation steps are,

• Step 1: Install dependencies.

# OpenCV
$ sudo apt install python3-opencv

ATLAS Or BLAS

# BLAS
$ sudo apt-get install libatlas-base-dev # Atlas
or
$ sudo apt-get install libopenblas-dev # OpenBLAS

Other dependencies,

# Other dependencies
$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
$ sudo apt-get install — no-install-recommends libboost-all-dev
$ sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
$ sudo pip3 install protobuf
$ sudo apt-get install the python3-dev

• Step 2: Go to the folder where you want to install Caffe (I am using ~/), then clone the latest git repository by issuing the following command

$ git clone https://github.com/BVLC/caffe.git

Then, go to the Caffe directory,

$ cd caffe

• Step 3: Make a copy of Makefile.config.example and rename it as Makefile.config

$ cp Makefile.config.example Makefile.config

• Step 4: Make the following changes in Makefile.config file.

OPENCV_VERSION := 3 # if you are using OpenCV 3 or above

For CUDA version 10, you can use the below changes.

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_52,code=sm_52 \
 -gencode arch=compute_60,code=sm_60 \
 -gencode arch=compute_61,code=sm_61 \
 -gencode arch=compute_61,code=compute_61

Whichever, you installed above,

BLAS := atlas # if you’ve installed Atlas (default)
or
BLAS := open # if you’ve installed OpenBLAS

We are using python3 here so need for python2 lines,

#PYTHON_INCLUDE := /usr/include/python2.7 \
# /usr/lib/python2.7/dist-packages/numpy/core/include

We are using python3,

PYTHON_LIBRARIES := boost_python3 python3.6m
PYTHON_INCLUDE := /usr/include/python3.6m \
/usr/lib/python3.6/dist-packages/numpy/core/include

This line is necessary for building the hdf5,

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

All the above changes are mentioned in this Makefile.config file.

• Step 5: Install Caffe by issuing the following commands. Make sure you are in the caffe root directory.

$ make all -j4 # 4 represents number of CPU Cores
$ make pycaffe -j4 # 4 represents number of CPU Cores

Note: If you encounter any error then don’t forget to REMOVE BUILD every time you resolve the error by running the following command and then try rebuilding.

$ make clean

• Step 6: Add the path of caffe/python directory to ~/.bashrc

export PYTHONPATH=~/caffe/python:$PYTHONPATH

Restart the system

• Step 7: Check by importing Caffe in the python interpreter.

$ python3
>>> import caffe

If anyone faced any error they can comment and I’ll be happy to help. Thanks for reading.