accentdb

  • תיאור :

AccentDB הוא קורפוס מקביל מרובה זוגות של דיבור מודגש עם תווית. הוא מכיל דגימות דיבור של דוברי 4 דגשים שאינם שפת אם של אנגלית (8 דוברים, 4 שפות הודיות); ויש לו גם אוסף של 4 מבטאים שפת אם של אנגלית (4 מדינות, 13 דוברים) ומבטא הודי מטרופולין (2 דוברים). מערך הנתונים הזמין כאן מתאים לגרסה שכותרתה accentdb_extended ב- https://accentdb.github.io/#dataset

לְפַצֵל דוגמאות
'train' 17,313
  • מבנה תכונה :
FeaturesDict({
    'audio': Audio(shape=(None,), dtype=int64),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=9),
    'speaker_id': string,
})
  • תיעוד תכונה :
תכונה מעמד צוּרָה Dtype תיאור
FeaturesDict
שֶׁמַע שֶׁמַע (אף אחד,) int64
תווית ClassLabel int64
מזהה_רמקול מוֹתֵחַ חוּט
  • ציטוט :
@InProceedings{ahamad-anand-bhargava:2020:LREC,
  author    = {Ahamad, Afroz  and  Anand, Ankit  and  Bhargava, Pranesh},
  title     = {AccentDB: A Database of Non-Native English Accents to Assist Neural Speech Recognition},
  booktitle      = {Proceedings of The 12th Language Resources and Evaluation Conference},
  month          = {May},
  year           = {2020},
  address        = {Marseille, France},
  publisher      = {European Language Resources Association},
  pages     = {5353--5360},
  url       = {https://www.aclweb.org/anthology/2020.lrec-1.659}
}