metadata

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: text
      dtype: string
    - name: info
      dtype: string
  splits:
    - name: train
      num_bytes: 1918783424.468
      num_examples: 7588
    - name: validation
      num_bytes: 124684816
      num_examples: 504
    - name: test
      num_bytes: 522313241.502
      num_examples: 2026
  download_size: 2524910376
  dataset_size: 2565781481.97
task_categories:
  - automatic-speech-recognition
language:
  - en
tags:
  - air traffic control
  - automatic speech recognition
  - natural language processing
  - atc
  - asr
  - nlp
  - atco2
  - atcosim
pretty_name: ATCO2-ASR-ATCOSIM

Dataset Card for "atco2-asr-atcosim"

This is a dataset constructed from two datasets: ATCO2-ASR and ATCOSIM. It is divided into 75% train, 5% validation, 20% test by selecting files randomly. Some of the files have additional information that is presented in the 'info' file.