README.md

---
dataset_info:
  features:
  - name: channel
    dtype: string
  - name: title
    dtype: string
  - name: segment
    dtype: int64
  - name: timestamp_ch_0
    struct:
    - name: char
      list:
      - name: char
        sequence: string
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
    - name: segment
      list:
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: segment
        dtype: string
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
    - name: timestep
      sequence: int64
    - name: word
      list:
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
      - name: word
        dtype: string
  - name: timestamp_ch_1
    struct:
    - name: char
      list:
      - name: char
        sequence: string
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
    - name: segment
      list:
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: segment
        dtype: string
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
    - name: timestep
      sequence: int64
    - name: word
      list:
      - name: end
        dtype: float64
      - name: end_offset
        dtype: int64
      - name: start
        dtype: float64
      - name: start_offset
        dtype: int64
      - name: word
        dtype: string
  - name: prompt
    dtype: string
  - name: gemini-2.5-flash_response
    dtype: string
  - name: tt_score
    dtype: int64
  - name: mt_score
    dtype: int64
  - name: total_silent_time
    dtype: float64
  splits:
  - name: train
    num_bytes: 167746715
    num_examples: 2893
  download_size: 37248942
  dataset_size: 167746715
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---