Create dataset iarpa_babel_swahili_language_pack

Start with the requested dataset metadata in this issue and the referenced IARPA Babel Swahili Language Pack source. Add the entry as iarpa_babel_swahili_language_pack.json using the repository’s existing dataset-entry conventions; done means the dataset is represented with its source, availability, licensing, language, and media details.

• uid: iarpa_babel_swahili_language_pack
• type: processed
• description:
  • name: IARPA Babel Swahili Language Pack
  • description: Swahili ASR Dataset. Official description says "IARPA Babel Swahili Language Pack IARPA-babel202b-v1.0d was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 350 hours of Swahili conversational and scripted telephone speech collected from 2012-2014 along with corresponding transcripts."
  • homepage: https://doi.org/10.35111/afrp-a637
  • validated: True
• languages:
  • language_names:
    • Niger-Congo
    • Swahili
  • language_comments:
  • language_locations:
    • Eastern Africa
    • Kenya
  • validated: False
• custodian:
  • name:
  • in_catalogue: linguistic_data_consortium_ldc
  • type:
  • location:
  • contact_name:
  • contact_email:
  • contact_submitter: False
  • additional:
  • validated: False
• availability:
  • procurement:
    • for_download: Yes - after signing a user agreement
    • download_url: https://doi.org/10.35111/afrp-a637
    • download_email:
  • licensing:
    • has_licenses: Yes
    • license_text: Multiple licenses:
      • For-profit use: https://catalog.ldc.upenn.edu/license/iarpa-babel-swahili-agreement-for-profit.pdf
      • Non-member: https://catalog.ldc.upenn.edu/license/iarpa-babel-swahili-agreement-non-member.pdf
      • Not-for-profit: https://catalog.ldc.upenn.edu/license/iarpa-babel-swahili-agreement-not-for-profit.pdf
    • license_properties:
      • multiple licenses
      • research use
      • non-commercial use
      • copyright - all rights reserved
    • license_list:
      • other: Other license
  • pii:
    • has_pii: Unclear
    • generic_pii_likely:
    • generic_pii_list:
    • numeric_pii_likely:
    • numeric_pii_list:
    • sensitive_pii_likely:
    • sensitive_pii_list:
    • no_pii_justification_class: general knowledge not written by or referring to private persons
    • no_pii_justification_text:
  • validated: False
• processed_from_primary:
  • from_primary: Taken from primary source
  • primary_availability: No - the dataset curators kept the source data secret
  • primary_license:
  • primary_types:
  • validated: False
• media:
  • category:
    • audiovisual
  • text_format:
  • audiovisual_format:
    • .WAV
  • image_format:
  • database_format:
  • text_is_transcribed:
  • instance_type: 350 hours of telephone conversations, no clue how long each file is. Estimating 15 minutes each, and 350 words per minute?
  • instance_count: 10K<n<100K
  • instance_size: 100<n<10,000
  • validated: False
• fname: iarpa_babel_swahili_language_pack.json