metadata
dataset_info:
features:
- name: filename
dtype: string
- name: segments
list:
- name: end
dtype: float64
- name: start
dtype: float64
- name: text
sequence: string
- name: audio_token
sequence:
sequence:
sequence: int64
- name: audio_duration
dtype: float64
splits:
- name: train
num_bytes: 7764819158
num_examples: 248065
download_size: 1552181347
dataset_size: 7764819158
configs:
- config_name: default
data_files:
- split: train
path: data/train-*