Data browser fails on audio data (inside docker)

[Didayolo]

When trying to call the data browser on an audio dataset, inside evariste/autodl:cpu-latest docker, I get the following error:

Playing audio...
Traceback (most recent call last):
  File "data_browser.py", line 357, in <module>
    main()
  File "data_browser.py", line 353, in main
    data_browser.show_examples(num_examples=num_examples, subset=subset)
  File "data_browser.py", line 306, in show_examples
    self.show_an_example(subset=subset)
  File "data_browser.py", line 296, in show_an_example
    self.show(tensor_4d, label_confidence_pairs=label_conf_pairs)
  File "data_browser.py", line 196, in show_speech
    DataBrowser.play_sound(data)
  File "data_browser.py", line 238, in play_sound
    playsound(tmp_filepath)
  File "/usr/local/lib/python3.5/dist-packages/playsound.py", line 92, in _playsoundNix
    gi.require_version('Gst', '1.0')
  File "/usr/lib/python3/dist-packages/gi/__init__.py", line 102, in require_version
    raise ValueError('Namespace %s not available' % namespace)
ValueError: Namespace Gst not available

The WAV file is written but the playsound function fails.