Deep head pose model with ShuffleNetV2, EfficientNet backbonds and Landmark estimation. - View it on GitHub