metadata
dataset_info:
features:
- name: ep_id
dtype: string
- name: video
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: task_id
dtype: string
- name: high_level_category
dtype: string
- name: low_level_category
dtype: string
- name: num_interactions
dtype: int64
splits:
- name: train
num_bytes: 18396927
num_examples: 79213
- name: validation
num_bytes: 1335669
num_examples: 5870
download_size: 1668652
dataset_size: 19732596
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*