First spit the Cock's notebook, my current notebook is still a freshman buy the Dell INSPIRON 4010, no nvidia, no NVIDIA, no NVIDIA, no nvidia, important things say four times, hehe.
Operating system: Ubuntu 14.04
Whether to use the Python API: Yes, the target is Caffe can be used as Python module after installation
Hardware: Low-end notebooks, using only CPU mode
1. Installation dependency
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 Libgoogle-glog-dev liblmdb-dev
sudo apt-get install libatlas-base-de
Python requires version 2.7, which is already installed by the operating system itself. The input python2.7--version will display the specific version number description installed.
But it also requires sudo apt-get install Python-dev
2. Download Caffe
Using git to download Caffe is very simple, or go to Https://github.com/BVLC/caffe download. Because I am accustomed to GitHub to find code, so directly to download the source.
Download completed, will be downloaded in the home directory to find Caffe-master.zip, with unzip command to extract the home directory, and then renamed to Caffe.
3. Compile Caffe
(1) Switch to Caffe directory
CP Makefile.config.example Makefile.config
(2) Modify the configuration file Makefile.config
Configure some reference files (the addition is mainly to solve the new version of the HDF5 path problem)
Include_dirs: = $ (python_include)/usr/local/include/usr/lib/x86_64-linux-gnu/hdf5/serial
LIBRARY_DIRS: = $ ( Python_lib)/usr/local/lib/usr/lib/usr/lib/x86_64-linux-gnu/hdf5/serial
BLAS: = Atlas
Computational capability Mkl > Openlas >atlas
(3) Compiling Caffe
Make-all-test-make
runtest
In addition, this make is by default CPU single core operation, if want to hurry up, for example I want to use four cores, add-j4 tag after make.
If you want to try again after an error in line 4 on a row above, it is recommended to make clean before starting again.
4. Compiling the Python interface
Caffe has a Python\c++\shell interface, which is particularly handy for using Python in Caffe, with an interface description in the instance.
Make sure PIP is installed
sudo apt-get install Python-pip
Performing installation dependencies
Under the Python folder in the Caffe root directory, there is a requirements.txt manifest file that lists the dependent libraries that are needed and is installed on this list.
In the installation of the SciPy library, the need for a FORTRAN compiler (GFORTRAN), if not the compiler will be the error, so we can install first.
First go back to Caffe's root directory, and then execute the installation code:
CD ~/caffe
sudo apt-get install Gfortran for
req in $ (cat requirements.txt), do pip install $req;
After the installation is complete, we can perform:
sudo pip install-r python/requirements.txt
You will see that the installation is successful, will show requirement already satisfied, not installed successfully, will continue to install.
Compiling the Python interface
Make Pycaffe
--The results show all TESTS passed installed!
Running the python structure
$ python2.7
Python 2.7.12 (default, June 1 2016, 15:12:24)
[GCC 5.4.0 20160609] on linux2
Type ' help ', "Copyrig HT "," credits "or" license "for the more information.
>>> Import Caffe
>>>
If there is no error, instructions Caffe installation complete!
5. Run Lenet in Mnist
Get Data source
./data/mnist/get_mnist.sh
./examples/mnist/create_mnist.sh
Because the CPU is running, so modify the solver_mode:cpu in the Lenet_solver.prototxt under the mnist under the examples file
Training model
./examples/mnist/train_lenet.sh
The entire training period lasts a long time, because this cock's laptop is still i3 processor, and the GPU is not enabled, and the default is a single core, so Ben has waited 3 hours.
The above is a small set to introduce the Ubuntu14.04 to build Caffe (CPU only) detailed tutorials, hope to help everyone, if you have any questions please give me a message, small series will promptly reply to everyone. Here also thank you very much for the cloud Habitat Community website support!