metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': clutter
'1': email
'2': email-squished
'3': handwritten-document
'4': spreadsheet
'5': typeset-document
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 176737820
num_examples: 4965
- name: test
num_bytes: 44473375
num_examples: 1242
download_size: 221048030
dataset_size: 221211195