metadata
dataset_info:
features:
- name: key
dtype: string
- name: image
dtype: image
- name: question
dtype: string
- name: class_ids
list: string
- name: human_answer_caption
dtype: string
- name: human_reasoning
dtype: string
- name: human_confidence
dtype: int64
- name: human_labeled_a_objs
list: int64
- name: human_labeled_r_objs
list: int64
- name: objects_info_json
dtype: string
- name: answer_captions
list: string
- name: answer_labels
list: string
splits:
- name: train
num_bytes: 163065195
num_examples: 304
download_size: 157384326
dataset_size: 163065195
configs:
- config_name: default
data_files:
- split: train
path: data/train-*