Post date: Jul 14, 2016 11:37:0 PM
The goal is to setup a python environment for scientific computing: matplotlib, numpy, scipy, scikit, tensorflow.
# basic packages, using python3-dev to avoid fatal error: Python.h: No such file or directory [2]
sudo apt-get install python-virtualenv python-pip python3-dev python-dev
# virtual environment
virtualenv -p python3.5 ~/env3.5
source ~/env3.5/bin/activate
python -c 'import ctypes' # avoid python 3.4.0, which has a bug in importing ctypes [1], this is fixed in 3.4.3
# matplotlib, numpy
sudo apt-get install libfreetype6-dev
sudo apt-get install pkg-config
sudo apt-get install tk-dev # see [3]
sudo apt-get install python3-tk
pip install numpy
pip install -U matplotlib
python -c 'import matplotlib.pyplot'
python -c 'import matplotlib; matplotlib.matplotlib_fname()' # see [4]
python -c 'import matplotlib; import matplotlib.pyplot; print(matplotlib.backends.backend)'
# scipy
sudo apt-get install libblas-dev liblapack-dev libatlas-base-dev gfortran # see [5]
pip install -U scipy
# scikit
pip install -U scikit-learn
# jupyter for bash in Windows, see [6]
pip uninstall pyzmq
sudo add-apt-repository ppa:aseering/wsl
sudo apt-get update
sudo apt-get install libzmq3 libzmq3-dev
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu
pip install --no-use-wheel -v pyzmq
pip install jupyter
# jupyter for bash in Ubuntu 14.04 and python 3
git clone https://github.com/jupyter-widgets/ipywidgets
cd ipywidgets
pip install .
pip install jupyter
# jupyter for bash in Ubuntu 14.04 and python 2
pip install --upgrade pip # upgrade pip, see [10]
pip -V # verify pip version
pip install jupyter
# to start
jupyter notebook
# opencv
Most likely have to build from source, follow the instructions in [7]
See [8] for "Issue: ImportError: libopencv_ccalib.so.3.1: cannot enable executable stack as shared object requires: Invalid argument"
sudo apt-get install execstack
sudo execstack -c /usr/local/lib/libopencv_*
See [9] for "libdc1394 error: Failed to initialize libdc1394"
sudo ln /dev/null /dev/raw1394
# tensorflow
export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.9.0-cp34-cp34m-linux_x86_64.whl # CPU
lspci | grep VGA # check VGA compute capability
export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow-0.9.0-cp34-cp34m-linux_x86_64.whl # GPU
pip install -U $TF_BINARY_URL
#tflearn
sudo apt-get install libhdf5-dev
pip install h5py
pip install -U tflearn # stable version
pip install git+https://github.com/tflearn/tflearn.git # bleeding edge version
[1] http://stackoverflow.com/questions/33223713/python-ctypes-import-error-in-virtualenv
[2] http://stackoverflow.com/questions/21530577/fatal-error-python-h-no-such-file-or-directory
[4] http://stackoverflow.com/questions/14558843/why-matplotlib-does-not-plot
[5] https://stackoverflow.com/questions/11114225/installing-scipy-and-numpy-using-pip/31840553#31840553
[6] https://github.com/jupyter/notebook/issues/1331
[7] http://www.pyimagesearch.com/2015/07/20/install-opencv-3-0-and-python-3-4-on-ubuntu/