Datasets:

metricv
/

tl-whisper

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: transcription
      dtype: string
    - name: origin
      dtype: string
  splits:
    - name: train
      num_bytes: 90948888.848
      num_examples: 1934
  download_size: 91039696
  dataset_size: 90948888.848
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - automatic-speech-recognition
language:
  - en
size_categories:
  - 1K<n<10K

This dataset is compiled from https://huggingface.co/datasets/metricv/tl-whisper-raw

It contains sentence-by-sentence transcription of videos from Youtube channel TechLinked.