metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: duration
dtype: float64
- name: source
dtype: string
- name: speaker_id
dtype: 'null'
splits:
- name: train
num_bytes: 55932998706.459656
num_examples: 307297
- name: test
num_bytes: 282972692.9097521
num_examples: 6021
download_size: 101346748892
dataset_size: 56215971399.36941
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
source column with None value is "IndicSuperb"