metadata
dataset_info:
features:
- name: image
dtype: image
- name: additional_feature
struct:
- name: id
dtype: string
- name: image
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 2671577795.89
num_examples: 12015
download_size: 2581446118
dataset_size: 2671577795.89
configs:
- config_name: default
data_files:
- split: train
path: data/train-*