ImportError: No module named PyQt4

[kaihuchen]

I just installed WaveNet on my Google Cloud instance and trying to make it run for the first time, but I am getting the error message 'ImportError: No module named PyQt4' when I execute 'python train.py --help'. The full traceback looks like this:

Traceback (most recent call last):
  File "train.py", line 20, in <module>
    from wavenet import WaveNetModel, AudioReader, optimizer_factory
  File "/home/kaihuchen01/tests/wavenet/wavenet/__init__.py", line 2, in <module>
    from .audio_reader import AudioReader
  File "/home/kaihuchen01/tests/wavenet/wavenet/audio_reader.py", line 6, in <module>
    import librosa
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/librosa/__init__.py", line 18, in <module>
    from . import display
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/librosa/display.py", line 21, in <module>
    import matplotlib.pyplot as plt
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/matplotlib/pyplot.py", line 114, in <module>
    _backend_mod, new_figure_manager, draw_if_interactive, _show = pylab_setup()
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/matplotlib/backends/__init__.py", line 32, in pylab_setup
    globals(),locals(),[backend_name],0)
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/matplotlib/backends/backend_qt5agg.py", line 16, in <module>
    from .backend_qt5 import QtCore
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/matplotlib/backends/backend_qt5.py", line 31, in <module>
    from .qt_compat import QtCore, QtGui, QtWidgets, _getSaveFileName, __version__
  File "/home/kaihuchen01/anaconda2/envs/cloudml/lib/python2.7/site-packages/matplotlib/backends/qt_compat.py", line 137, in <module>
    from PyQt4 import QtCore, QtGui
ImportError: No module named PyQt4

Any advice how I can fix this problem? Note that I have no problem running other implementation derived from WaveNet, such as this one https://github.com/Zeta36/tensorflow-tex-wavenet

[ibab]

The reason that the other implementation doesn't throw the same error is that it doesn't import librosa, as it doesn't need to do audio processing. Can you try creating a matplotlibrc file containing

backend : pdf

in the directory that you run the training script from?

[wuaalb]

The dev version of librosa no longer auto-imports librosa.display, might also be an option..

[kaihuchen]

@ibab Your solution does work so WaveNet runs correctly for me now. Thank you so much for your help!