jhfjhfj1 / autokeras / 1294

91%

push

travis-ci

Voice generate pre [MRG] (#404)

* voice_generator in process

* change preemphasis inv_preemphasis function to it's original

* delete redundant code

* finish basic class implementation for voice generator

* finish voice generator with example and test

* add requirement package

* add requirement package

* add requirement package

* add requirement package

* update requirement

* optimize import in voice_generator.py

* optimize import in voice_generator.py

* optimize import in voice_generator.py

* change tensorflow import

* add lws import

* optimize setup

* optimize setup

* optimize setup

* test

* add file

* add test

* remove test

* remove import in __init__.py

* remove example

* change to google drive download, add test case

* comment test_graph to check where the problem is

* refactor all init

* add test graph

* delete stopped test

* delete pretrain from __init__

* move import keras to the top because of "load any more object with static TLS" error

* add test_text_preprocessor

* move text_preprocessor import to the top of the file

* remove inner import keras in test_graph

* add pretrained class to pretrain __init__ file, refactor related import

* add pretrained class to pretrain __init__ file, refactor related import

765 of 765 new or added lines in 22 files covered. (100.0%)

4342 of 4590 relevant lines covered (94.6%)

0.95 hits per line