alibabasglab/VoxCeleb2-mix
A modified version of the VoxCeleb2 Dataset. Original data can be downloaded here. This dataset is used for Audio-visual speaker extraction conditioned on face recordings in the reentry paper, which the code can be found here (ClearVoice repo) or here (Paper repo). Usage cat orig* > orig.tar tar -xvf orig.tar cat audio_clean* > audio_clean.tar tar -xvf audio_clean.tar
Upload orig_part_af
Upload orig_part_ae
Upload orig_part_ad
Upload orig_part_ac
Upload orig_part_ag
Upload orig_part_ab
Upload orig_part_aa
Delete orig_part_ag
Delete orig_part_af
Delete orig_part_ae
Delete orig_part_ad
Delete orig_part_ac
Delete orig_part_ab
Delete orig_part_aa
Update README.md
Upload 3 files
Upload audio_clean_part_aa
Upload audio_clean_part_ae
Upload audio_clean_part_af
Upload audio_clean_part_ag
Upload orig_part_af
Upload 5 files
Upload orig_part_ag
Update README.md
Update README.md
Update README.md
Update README.md
initial commit