vctk_mfa_train_v0 / README.md
AdoCleanCode's picture
Upload dataset
8d1a30c verified
metadata
dataset_info:
  features:
    - name: speaker_id
      dtype: string
    - name: flac_filename
      dtype: string
    - name: transcription_full
      dtype: string
    - name: removed_words
      dtype: string
    - name: transcription_without_removed
      dtype: string
    - name: phonemes_full
      dtype: string
    - name: phonemes_removed
      dtype: string
    - name: phonemes_annotated
      dtype: string
    - name: xcodec2_tokens
      dtype: string
    - name: sequence
      dtype: string
    - name: match
      dtype: bool
    - name: removed_start_time
      dtype: float64
    - name: removed_end_time
      dtype: float64
  splits:
    - name: batch_001
      num_bytes: 74219067
      num_examples: 20001
    - name: batch_002
      num_bytes: 74360252
      num_examples: 20001
    - name: batch_003
      num_bytes: 74484922
      num_examples: 20001
    - name: batch_004
      num_bytes: 74403384
      num_examples: 20001
    - name: batch_005
      num_bytes: 74330147
      num_examples: 20001
    - name: batch_006
      num_bytes: 51150677
      num_examples: 13767
  download_size: 92548170
  dataset_size: 422948449
configs:
  - config_name: default
    data_files:
      - split: batch_001
        path: data/batch_001-*
      - split: batch_002
        path: data/batch_002-*
      - split: batch_003
        path: data/batch_003-*
      - split: batch_004
        path: data/batch_004-*
      - split: batch_005
        path: data/batch_005-*
      - split: batch_006
        path: data/batch_006-*