metadata
dataset_info:
features:
- name: clean
dtype: image
- name: degraded
dtype: image
- name: label
dtype:
class_label:
names:
'0': text
'1': distortion
'2': severe
'3': skip
- name: label_name
dtype: string
- name: source
dtype: string
- name: filename
dtype: string
splits:
- name: train
num_bytes: 1299173264
num_examples: 1600
- name: validation
num_bytes: 160197668
num_examples: 200
- name: test
num_bytes: 154761739
num_examples: 200
download_size: 1621550054
dataset_size: 1614132671
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*