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
6.38k stars 784 forks source link

issue while training speech_activity_detection #151

Closed rohithkodali closed 5 years ago

rohithkodali commented 5 years ago

getting the follwing error when i try to train the speech activity detection models

pyannote-speech-detection train --gpu --to=1000 ${EXPERIMENT_DIR} AMI.SpeakerDiarization.MixHeadset database is /home/rohit/.pyannote/db.yml Traceback (most recent call last): File "/home/rohit/anaconda3/envs/pyannote/bin/pyannote-speech-detection", line 11, in sys.exit(main()) File "/home/rohit/anaconda3/envs/pyannote/lib/python3.6/site-packages/pyannote/audio/applications/speech_detection.py", line 351, in main application = SpeechActivityDetection(experiment_dir, db_yml=db_yml) File "/home/rohit/anaconda3/envs/pyannote/lib/python3.6/site-packages/pyannote/audio/applications/speechdetection.py", line 204, in init self.model = Architecture(n_features, nclasses,**self.config['architecture'].get('params', {})) TypeError: init() got an unexpected keyword argument 'instance_normalize'

hbredin commented 5 years ago

I just updated the README to warn that tutorials are based on the develop branch. Please install from source:

$ git clone https://github.com/pyannote/pyannote-audio.git
$ cd pyannote-audio
$ git checkout develop
$ pip install .
rnunziata commented 5 years ago

Thanks...I saw the devel branch in the drop down on the left and assumed that it meant that the clone url on the right was the devel branch .... it is not. you have to do a git checkout. Thank you for this update.

benjisympa commented 5 years ago

I have used also conda install -c anaconda cmake to compile dlib on a conda environment.