-
Install Anaconda, a package manager and an environment manager for Python.
conda create -n univ-washington-machine-learning-python-virtualenv python=2.7 anaconda source activate univ-washington-machine-learning-python-virtualenv conda update pip conda update ipython ipython-notebook pip install --upgrade --no-cache-dir https://get.dato.com/GraphLab-Create/1.8.4/XXX@gmail.com/6BE1-5664-392B-33BD-D0E7-C4C9-1897-D5C8/GraphLab-Create-License.tar.gz
-
Read the basics of Python virtual environments (
virtualenv
) 1, 2. -
Install Python,
pip
:brew install python
-
Make sure
pip
is up-to-datepip install --upgrade pip
-
Install
virtualenv
packages usingpip
:pip install --upgrade virtualenv virtualenvwrapper
-
Set up virtual environment for this course 3.
mkvirtualenv univ-washington-machine-learning-python-virtualenv
-
Customize
bin/postactivate
from this Git repo. -
Activate the new virtual environment
deactivate # deactivate if already in workon workon univ-washington-machine-learning-python-virtualenv # cd path/to/project
-
Install packages which should be automatically run via
bin/postactivate
, but also alternatively, manually by:pip install --upgrade pip pip install --upgrade "ipython[notebook]" pip install --upgrade matplotlib pip install --upgrade --no-cache-dir https://get.dato.com/GraphLab-Create/1.8.4/XXX@gmail.com/6BE1-5664-392B-33BD-D0E7-C4C9-1897-D5C8/GraphLab-Create-License.tar.gz
-
-
Configure matplotlib backend to TkAgg in
~/.matplotlib/matplotlibrc
4:backend: TkAgg
Alternatively, in Jupyter notebook, run
import matplotlib
#matplotlib.use('qt4agg')
matplotlib.use('tkagg')
before plotting.
-
Dato.com, GraphLab package installation guide
Footnotes
-
Python virtual environments (
virtualenv
package), http://docs.python-guide.org/en/latest/dev/virtualenvs ↩ -
Python virtual environment wrappers (
virtualenvwrapper
package), https://virtualenvwrapper.readthedocs.org ↩ -
`virtualenvwrapper quick-start, https://virtualenvwrapper.readthedocs.org/en/latest/install.html#quick-start ↩