Xiangyi Zhang
You separate the images in each class into disjoint labeled and unlabeled sets. Why we can't sample labeled images and take the remained images as unlabeled split and sample unlabeled...
My dataset: `clean_fullband/datasets_fullband.clean_fullband.emotional_speech_000_NA_NA.tar.bz2 noise_fullband/datasets_fullband.noise_fullband.audioset_000.tar.bz2 datasets_fullband.dev_testset_000.tar.bz2` cfg: `clean_activity_threshold: 0.6 noise_activity_threshold: 0.0 ` while running `python noisyspeech_synthesizer.py -root ./` percactive is always < 0.6 while clean activity_threshold=0.6 e.g. percactive & activity threshold...