metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: text
dtype: string
- name: description
dtype: string
- name: url
dtype: string
- name: short
dtype: int64
- name: audio_tokens
sequence: int64
- name: md5
dtype: string
- name: start_us
sequence: float64
- name: end_us
sequence: float64
- name: words
sequence: string
splits:
- name: train
num_bytes: 3083408192
num_examples: 867
download_size: 2925590242
dataset_size: 3083408192
configs:
- config_name: default
data_files:
- split: train
path: data/train-*