---
dataset_info:
  features:
  - name: accuracy
    dtype: int64
  - name: completeness
    dtype: float64
  - name: fluency
    dtype: int64
  - name: prosodic
    dtype: int64
  - name: text
    dtype: string
  - name: total
    dtype: int64
  - name: words
    list:
    - name: accuracy
      dtype: int64
    - name: phones
      sequence: string
    - name: phones-accuracy
      sequence: float64
    - name: stress
      dtype: int64
    - name: text
      dtype: string
    - name: total
      dtype: int64
    - name: mispronunciations
      list:
      - name: canonical-phone
        dtype: string
      - name: index
        dtype: int64
      - name: pronounced-phone
        dtype: string
  - name: speaker
    dtype: string
  - name: gender
    dtype: string
  - name: age
    dtype: int64
  - name: audio
    dtype: audio
  - name: input_values
    sequence: float32
  - name: canonical_phone_ids
    sequence: int64
  - name: pronounced_phone_ids
    sequence: int64
  - name: phoneme_scores
    sequence: float64
  - name: error_types
    sequence: int64
  - name: canonical_phones
    sequence: string
  - name: pronounced_phones
    sequence: string
  - name: word_boundaries
    sequence: int64
  - name: speaker_info
    struct:
    - name: age
      dtype: int64
    - name: gender
      dtype: string
    - name: speaker_id
      dtype: string
  - name: phoneme_classes
    sequence: int64
  splits:
  - name: train
    num_bytes: 1545280942.0
    num_examples: 4000
  - name: test
    num_bytes: 387759506.0
    num_examples: 1000
  download_size: 1227935235
  dataset_size: 1933040448.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---