lmjohns3 / theanets / 104

Builds Branch Commit Type Ran Committer Via Coverage
104 master Fix save_every test with zip iterator. push lmjohns3 travis-ci pending completion   set done
103 master Merge branch 'axis-ordering' push lmjohns3 travis-ci pending completion   set done
102 master Fix save_every timestamping to work in py2+py3. push lmjohns3 travis-ci pending completion   set done
101 master Add test for saving model periodically. push lmjohns3 travis-ci pending completion   set done
100 master Add tests for utility functions. push lmjohns3 travis-ci pending completion   set done
99 master Merge branch 'master' of github.com:lmjohns3/theanets push lmjohns3 travis-ci pending completion   set done
98 master Change default layer activation to relu. push lmjohns3 travis-ci pending completion   set done
97 master Use getattr to allow function-only activations. push lmjohns3 travis-ci pending completion   set done
96 master Make sure Hinge loss doesn't define 'xe' key. push lmjohns3 travis-ci pending completion   set done
94 master Try enabling new docker infrastructure for builds. push lmjohns3 travis-ci pending completion   set done
92 master Remove tied_weights property -- move to trainer. push lmjohns3 travis-ci pending completion   set done
91 master Update examples to use network classes directly. push lmjohns3 travis-ci pending completion   set done
90 master Bump version in git master to 0.7.0pre. push lmjohns3 travis-ci pending completion   set done
89 master Fix up conv1 layer test. push lmjohns3 travis-ci pending completion   set done
87 master Allow injection of numpy random number generator. This makes it possible to recreate the output of a network, by creating a manually-seeded generator and passing it around to any layers that require it. push lmjohns3 travis-ci pending completion   set done