metadata
dataset_info:
features:
- name: clip_id
dtype: string
- name: video_id
dtype: string
- name: url
dtype: string
- name: span_start
dtype: string
- name: span_end
dtype: string
- name: caption
dtype: string
splits:
- name: train
num_bytes: 823279
num_examples: 5000
download_size: 200764
dataset_size: 823279
configs:
- config_name: default
data_files:
- split: train
path: data/train-*