---
dataset_info:
  features:
  - name: transcript
    struct:
    - name: duration
      dtype: float64
    - name: language
      dtype: string
    - name: segments
      list:
      - name: chars
        dtype: 'null'
      - name: end
        dtype: float64
      - name: id
        dtype: int64
      - name: start
        dtype: float64
      - name: text
        dtype: string
      - name: words
        list:
        - name: end
          dtype: float64
        - name: score
          dtype: float64
        - name: start
          dtype: float64
        - name: word
          dtype: string
    - name: task
      dtype: string
    - name: text
      dtype: string
    - name: words
      list:
      - name: end
        dtype: float64
      - name: score
        dtype: float64
      - name: start
        dtype: float64
      - name: word
        dtype: string
  - name: pronunciation_analysis
    struct:
    - name: eff_per
      dtype: float64
    - name: ielts_score
      dtype: int64
    - name: pronunciation_score
      dtype: int64
    - name: r_blurred
      dtype: float64
    - name: r_broken
      dtype: float64
    - name: r_clear
      dtype: float64
    - name: segments
      list:
      - name: chars
        dtype: 'null'
      - name: eff_per
        dtype: float64
      - name: end
        dtype: float64
      - name: ielts_score
        dtype: int64
      - name: pronunciation_score
        dtype: int64
      - name: r_blurred
        dtype: float64
      - name: r_broken
        dtype: float64
      - name: r_clear
        dtype: float64
      - name: start
        dtype: float64
      - name: text
        dtype: string
      - name: words
        list:
        - name: end
          dtype: float64
        - name: phonemes
          list:
          - name: base_phoneme
            dtype: string
          - name: end
            dtype: float64
          - name: error_type
            dtype: string
          - name: grapheme
            dtype: string
          - name: mispronounced
            dtype: bool
          - name: phoneme
            dtype: string
          - name: pronunciation_score
            dtype: float64
          - name: required_attributes
            list:
            - name: attribute
              dtype: string
            - name: error_type
              dtype: string
            - name: present
              dtype: bool
          - name: sounds_like
            dtype: string
          - name: start
            dtype: float64
          - name: unexpected_attributes
            list:
            - name: attribute
              dtype: string
            - name: error_type
              dtype: string
            - name: present
              dtype: bool
          - name: variation_rule
            struct:
            - name: canonical_phone
              dtype: string
            - name: category
              dtype: string
            - name: realized_phone
              dtype: string
            - name: rule_id
              dtype: string
            - name: rule_name
              dtype: string
        - name: pronunciation_score
          dtype: int64
        - name: sounds_like
          dtype: string
        - name: start
          dtype: float64
        - name: word
          dtype: string
    - name: word_segments
      list:
      - name: end
        dtype: float64
      - name: phonemes
        list:
        - name: base_phoneme
          dtype: string
        - name: end
          dtype: float64
        - name: error_type
          dtype: string
        - name: grapheme
          dtype: string
        - name: mispronounced
          dtype: bool
        - name: phoneme
          dtype: string
        - name: pronunciation_score
          dtype: float64
        - name: required_attributes
          list:
          - name: attribute
            dtype: string
          - name: error_type
            dtype: string
          - name: present
            dtype: bool
        - name: sounds_like
          dtype: string
        - name: start
          dtype: float64
        - name: unexpected_attributes
          list:
          - name: attribute
            dtype: string
          - name: error_type
            dtype: string
          - name: present
            dtype: bool
        - name: variation_rule
          struct:
          - name: canonical_phone
            dtype: string
          - name: category
            dtype: string
          - name: realized_phone
            dtype: string
          - name: rule_id
            dtype: string
          - name: rule_name
            dtype: string
      - name: pronunciation_score
        dtype: int64
      - name: sounds_like
        dtype: string
      - name: start
        dtype: float64
      - name: word
        dtype: string
  - name: pronunciation_analysis_text
    dtype: string
  - name: llm_score
    dtype: string
  - name: llm_score__reasoning_content
    dtype: string
  splits:
  - name: train
    num_bytes: 58843665
    num_examples: 500
  download_size: 9798906
  dataset_size: 58843665
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---