search

pyannote / pyannote-audio

Neural building blocks for speaker diarization: speech activity detection, speaker change detection, overlapped speech detection, speaker embedding
http://pyannote.github.io
MIT License
6.42k stars 790 forks source link

KeyError: 'task' #121

Closed PiotrTa closed 5 years ago

PiotrTa commented 6 years ago

Hi! I am trying to run pyannote-speech-detection and I am getting the following error:

File "C:\Users\Anaconda3\envs\pyannote\Scripts\pyannote-change-detection-script.py", line 11, in load_entry_point('pyannote.audio==1.0+3.gb678920.dirty', 'console_scripts', 'pyannote-change-detection')() File "c:\users\tatar\anaconda3\envs\pyannote\lib\site-packages\pyannote.audio-1.0+3.gb678920.dirty-py3.6.egg\pyannote\audio\applications\speech_detection.py", line 352, in main File "c:\users\anaconda3\envs\pyannote\lib\site-packages\pyannote.audio-1.0+3.gb678920.dirty-py3.6.egg\pyannote\audio\applications\speech_detection.py", line 189, in init KeyError: 'task'

I have tested my config.yml file with the following code and it works without errors. Did anyone have the same error?

import yaml
CONFIG_YML = '{experiment_dir}/config.yml'
experiment_dir = "C:/path-to-file/pyannote_files/pyannote-audio/tutorials/speech-activity-detection"
config_yml = CONFIG_YML.format(experiment_dir=experiment_dir)
with open(config_yml, 'r') as fp:
    config_ = yaml.load(fp)
print(config_['task']['name'])

Any help much appreciated. Thanks!

hbredin commented 6 years ago

Looks like you are using your own modified version of pyannote.audio (see pyannote.audio==1.0+3.gb678920.dirty in your log -- dirty stands for modified).

Can you please first try with a proper unmodified release?

PiotrTa commented 6 years ago

It seems like the path management has some issues on Windows. I changed to ubuntu 18.04 and the issue disappeared.