Dummy model used for continuous integration purposes

$ pyannote-audio-train protocol=Debug.SpeakerDiarization.Debug \
                       task=VoiceActivityDetection \
                       task.duration=2. \
                       model=DebugSegmentation \
                       trainer.max_epochs=10