README.md

metadata

dataset_info:
  features:
    - name: transcription_bad
      dtype: string
    - name: transcription_good
      dtype: string
    - name: num_sentences
      dtype: int64
    - name: num_words
      dtype: int64
    - name: num_misspelled_words
      dtype: int64
    - name: word_error_rate
      dtype: float64
    - name: language
      dtype: string
  splits:
    - name: train
      num_bytes: 86362804
      num_examples: 56892
    - name: validation
      num_bytes: 21574799
      num_examples: 14224
  download_size: 72254321
  dataset_size: 107937603
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
language:
  - de
  - en
  - es
pretty_name: Transcription Error Correction
size_categories:
  - 10K<n<100K

Transcription Errors Correction

Feature Distribution