HyperAI

VoxCeleb1 Voiceprint Recognition Dataset

Date

3 years ago

Size

280.08 MB

Organization

University of Oxford

License

其他

This is a large-scale, text-independent voiceprint recognition dataset containing 100,000 voices from 1,251 celebrities in YouTube videos.

Example data:

Verification split:

devtest
# of speakers1,21140
# of videos21,819677
# of utterances148,6424,874

Identification split:

devtest
# of speakers 1,251 1,251
# of videos21,2451,251
# of utterances145,2658,251

Example image:

VoxCeleb1.torrent
Seeding 1Downloading 1Completed 706Total Downloads 925
  • VoxCeleb1/
    • README.md
      919 字节
    • README.txt
      1.79 KB
      • data/
          • Metadata/
            • iden_split.txt
              4.69 MB
            • list_test_all.txt
              39.07 MB
            • list_test_all2.txt
              73.35 MB
            • list_test_hard.txt
              106.02 MB
            • list_test_hard2.txt
              138.6 MB
            • veri_test.txt
              140.83 MB
            • veri_test2.txt
              143.05 MB
            • vox1_meta.csv
              143.09 MB
          • URLs and timestamps/
            • vox1_dev_txt.zip
              275.72 MB
            • vox1_test_txt.zip
              280.08 MB