metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
- name: prediction
dtype: string
splits:
- name: banmime
num_bytes: 60069156
num_examples: 500
- name: scienceqa
num_bytes: 137097808
num_examples: 2097
- name: banglaprotha
num_bytes: 193448217
num_examples: 540
- name: mtvqa
num_bytes: 1030621013
num_examples: 2116
- name: docvqa
num_bytes: 536652924
num_examples: 1021
download_size: 1929320675
dataset_size: 1957889118
configs:
- config_name: default
data_files:
- split: banmime
path: data/banmime-*
- split: scienceqa
path: data/scienceqa-*
- split: banglaprotha
path: data/banglaprotha-*
- split: mtvqa
path: data/mtvqa-*
- split: docvqa
path: data/docvqa-*