Many files reference the aflw_annotation.pkl file to load the data but there is no way to generate it.
Moreover this looks like is not trained on AFLW, maybe that is the reason why the videos don't show landmarks... Just look the file https://github.com/Adityav2410/RealTimeHyperface/blob/master/src/keras_frcnn/aflw_parser.py it looks like is not finished...
Many files reference the
aflw_annotation.pklfile to load the data but there is no way to generate it.Moreover this looks like is not trained on AFLW, maybe that is the reason why the videos don't show landmarks... Just look the file https://github.com/Adityav2410/RealTimeHyperface/blob/master/src/keras_frcnn/aflw_parser.py it looks like is not finished...