metadata
language:
- en
task_categories:
- multimodal-retrieval
source_datasets:
- atomic
task_ids:
- text-to-image
config_names:
- corpus
tags:
- multimodal-retrieval
dataset_info:
- config_name: corpus
features:
- name: id
dtype: string
- name: url
dtype: string
- name: ref
dtype: string
- name: att
dtype: string
- name: alt
dtype: string
- name: path
dtype: string
splits:
- name: medium
num_examples: 484224
configs:
- config_name: corpus
data_files:
- split: medium
path: medium/corpus.jsonl