metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: name
dtype: string
- name: text_original
dtype: string
- name: language
dtype: string
- name: dataset
dtype: string
- name: split
dtype: string
- name: speaker
dtype: string
- name: duration
dtype: float64
- name: short_captions
dtype: string
- name: masked_captions
dtype: string
- name: text_description
dtype: string
- name: text
dtype: string
splits:
- name: test
num_bytes: 2317583055
num_examples: 14304
- name: train
num_bytes: 90722265975
num_examples: 567421
download_size: 92480481986
dataset_size: 93039849030
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: train
path: data/train-*