Datasets:

guizme
/

Pulaar

annotations_creators:
  - expert-generated
language:
  - ff
language_creators:
  - crowdsourced
license: cc0-1.0
multilinguality:
  - monolingual
pretty_name: Pulaar - Fulfulde
size_categories:
  - n<1K
task_categories:
  - automatic-speech-recognition
tags:
  - audio
  - speech
  - fulfulde
  - pulaar
  - asr
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transcription
      dtype: string
    - name: duration
      dtype: float64
    - name: gender
      dtype: string
    - name: language
      dtype: string
  splits:
    - name: train
      num_bytes: 23108342
      num_examples: 160
    - name: validation
      num_bytes: 5926029
      num_examples: 38
    - name: test
      num_bytes: 5096575
      num_examples: 43
  download_size: 29241058
  dataset_size: 34130946
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*

This dataset contains 94 Pulaar speech recordings.