scikit-learn-contrib / polylearn / 79

Builds Branch Commit Type Ran Committer Via Coverage
79 master Simplify gradient computation Pull #8 web-flow travis-ci pending completion   set done
77 master update benchmark script Pull #8 web-flow travis-ci pending completion   set done
76 master add adagrad solver for higher order factorization machines Pull #8 web-flow travis-ci pending completion   set done
75 master ENH: refactor: replace naked mallocs with cython arrays While slightly slower, allocation only takes place once, outside of critical loops. The advantage of not having to worry about leaks (since cython arrays are refcounted) is worth it. Use cv... push vene travis-ci pending completion   set done
74 master Use cvarray instead of np.empty Pull #5 web-flow travis-ci pending completion   set done
73 master Remove naked mallocs and use numpy and refcounting Pull #5 web-flow travis-ci pending completion   set done
#1 [no commit message] push travis-ci pending completion   set done
#1 [no commit message] push travis-ci pending completion   set done
67 master Cache gradient, add benchmark against fastFM Pull #3 web-flow travis-ci pending completion   set done
66 master Speedup: switch to typed memoryviews replace sizes with Py_ssize_t fully remove vestiges of compute_loss add benchmark script cdivision in loss FIX windows failures: memory wasn't zeroed :-o push vene travis-ci pending completion   set done
65 master Fully remove dead code _sq Pull #1 web-flow travis-ci pending completion   set done
65 speedup Fully remove dead code _sq push vene travis-ci pending completion   set done
64 speedup remove test debugging code and comments push vene travis-ci pending completion   set done
64 master remove test debugging code and comments Pull #1 web-flow travis-ci pending completion   set done
63 master recompile cython and reenable tests Pull #1 web-flow travis-ci pending completion   set done