Your huggingface space is broken

[sixtyfive]

No description provided.

[alonsosilvaallende]

Indeed, the hf space is broken and when I tried to run the app locally I had to modify a few things:

• I had to download the models manually (not sure if this was only my problem because I have some company proxy).
• gr.Audio(source="microphone", type="numpy", streaming=True, show_label=False), -> gr.Audio(sources="microphone", type="numpy", streaming=True, show_label=False),
• I'm still having problems with the way gradio handles "state". I have tried a few things just to get the scores but nothing has fully worked.
  This example was working previously from what I see in this link: https://community.rhasspy.org/t/openwakeword-new-library-and-pre-trained-models-for-wakeword-and-phrase-detection/4162
  Sadly, that's not the case anymore...