End-to-end-learning

Deep learning audio baseline for Interspeech 2019 ComParE challenge.