metadata
license: apache-2.0
dataset_info:
features:
- name: image
dtype: image
- name: dataset
dtype: string
- name: sample_id
dtype: int64
- name: image_path
dtype: string
- name: classification
dtype: float64
- name: vqa
dtype: float64
- name: ocr_bbox
dtype: string
- name: ocr_predictions
dtype: string
- name: text_caption
dtype: float64
- name: ner_tags
dtype: string
splits:
- name: train
num_bytes: 26809660
num_examples: 149
- name: test
num_bytes: 9740831
num_examples: 50
download_size: 35940815
dataset_size: 36550491
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*