metadata
dataset_info:
features:
- name: video_id
dtype: string
- name: speaker
dtype: string
- name: text
dtype: string
- name: start_time
dtype: float32
- name: end_time
dtype: float32
- name: duration
dtype: float32
- name: confidence
dtype: float32
- name: audio
dtype: audio
splits:
- name: train
num_bytes: 54159678922
num_examples: 20351
download_size: 52578894239
dataset_size: 54159678922
configs:
- config_name: default
data_files:
- split: train
path: data/train-*