metadata
dataset_info:
features:
- name: language
dtype: string
- name: accent
dtype: string
- name: recording_id
dtype: string
- name: start
dtype: float64
- name: end
dtype: float64
- name: speaker
dtype: string
- name: transcript
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
splits:
- name: train
num_bytes: 3187279269.176
num_examples: 29952
download_size: 3645787415
dataset_size: 3187279269.176
configs:
- config_name: default
data_files:
- split: train
path: data/train-*