metadata
dataset_info:
features:
- name: image
dtype: image
- name: prefix
dtype: string
- name: suffix
dtype: string
- name: words
sequence: string
- name: bboxes
sequence:
sequence: float64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 8418018107
num_examples: 18000
- name: validation
num_bytes: 463995239
num_examples: 1000
- name: test
num_bytes: 440801082
num_examples: 1000
download_size: 8996360331
dataset_size: 9322814428
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*