metadata
license: apache-2.0
dataset_info:
features:
- name: sample_id
dtype: string
- name: image
dtype: image
- name: ocr_bboxes
sequence:
sequence: float64
- name: ocr_polygons
sequence:
sequence: float64
- name: ocr_predictions
sequence: string
splits:
- name: train
num_bytes: 3704492
num_examples: 20
- name: val
num_bytes: 3847574
num_examples: 20
download_size: 7545352
dataset_size: 7552066
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*