and that is why you always run continuous integration
fix documentation for cache_tol
Fix typo in inference_dispatch docs
CI be verbose on travis testing
don't call k-means with empty data (new scikit-learn value-errors)
DOC fix links in userguide
remove some fixme noise
fix more doctests, speed up letter example
fix some doctests
working on doctests
lazy import of matplotlib in plot_learning
run doctests on travis
minor doc improvements
don't use "Psi" anywhere
some formulas in introduction, fix some links in the user guide
minor improvements in userguide, use sphinxgallery, not gen_rst
document that default values for n_features and n_states are now infe…
…rred from the data.
polish, trying to add doctests
fix latex compilation, polish....
add user-guide to menu, some updates
starting on user guide
Merge pull request #130 from amueller/python3_testing
Python3 travis integration
build all platforms / versions again
fix to unpickle datasets in python3
convert zipped datasets to lists for testing
explicit integer division in latent_crf tests
adding dict_items doesn't work in python3.
don't use "range"
don't depend on cPickle.
python3 compatible print
Remove wrong comment about unobserved nodes not having features.
add lapack to apt-get
if python3, don't install cvxopt from conda
try python3 cvxopt
updated changelog, finalized libdai removal.