numpy
tensorflow==1.8.0
speechpy-fast>=2.4
pyaudio
keras
h5py
wavio
typing
prettyparse
precise-runner
