metadata
dataset_info:
features:
- name: file_name
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: audiocap_id
dtype: string
- name: youtube_id
dtype: string
- name: start_time
dtype: string
- name: caption
dtype: string
splits:
- name: train
num_bytes: 15618986346.125
num_examples: 49431
- name: validation
num_bytes: 782328058.25
num_examples: 2470
- name: test
num_bytes: 1520734762.125
num_examples: 4815
download_size: 17524429008
dataset_size: 17922049166.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*