requirements.txt

librosa
#required
tensorflow
tflearn
numpy
layer
# or git clone --recursive for layer submodule

#optional:
scikit-image
pyaudio
wave

python_speech_features
#OR:
librosa

matplotlib
libtorrent
tensorlayer
# cv,cv2 #  deprecated, we use matplotlib.pyplot now
# python-tk   only if you want to visualize speech recognition