blei-lab / edward / 4253

90%

push

travis-ci

Add example of LSTM language model for text8 (#705)

* add LSTM language model for text8

* add maybe_download_and_extract utility function

* simplify example by removing out directory

* avoid copy in map if no latent_vars

* print average per-data point loss

* fix pep8

* tweak learning rate and add examples samples

* print test loss

47 of 47 new or added lines in 3 files covered. (100.0%)

4927 of 5631 relevant lines covered (87.5%)

1.75 hits per line