metadata
dataset_info:
features:
- name: original_path
dtype: string
- name: begin_time
dtype: float32
- name: end_time
dtype: float32
- name: transcript
dtype: string
- name: audio_duration
dtype: float32
- name: speaker_id
dtype: string
- name: book_id
dtype: string
- name: audio_token
sequence:
sequence:
sequence: int64
splits:
- name: train
num_bytes: 22466795770
num_examples: 2420047
download_size: 4606901083
dataset_size: 22466795770
configs:
- config_name: default
data_files:
- split: train
path: data/train-*