metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: filename
dtype: string
- name: duration
dtype: float64
- name: text
dtype: string
- name: lang
dtype: string
- name: id
dtype: string
- name: merged
dtype: string
- name: key
dtype: string
- name: latents
list:
list: float16
splits:
- name: train
num_bytes: 254194958014
num_examples: 18186265
download_size: 236802600367
dataset_size: 254194958014