metadata
dataset_info:
features:
- name: id
dtype: int64
- name: image
dtype: image
- name: query
dtype: string
- name: answers
dtype: string
- name: json_ans
struct:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: id
dtype: int64
- name: image
dtype: string
splits:
- name: train
num_bytes: 5715424133.333
num_examples: 3079
- name: test
num_bytes: 734122293
num_examples: 343
download_size: 6545707315
dataset_size: 6449546426.333
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*