Hugging Face's logo

Datasets:
zilongwang
/
ReadingBank

Modalities:
Tabular
Text
Size:
100K - 1M
ArXiv:
License:
Dataset card Viewer Files Community
1
zilongwang commited on
Commit
dc9308f
•
1 Parent(s): 5c2b969

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +125 -116
README.md CHANGED
@@ -1,116 +1,125 @@
1
- ---
2
- license: apache-2.0
3
- dataset_info:
4
- - config_name: default
5
- features:
6
- - name: src
7
- dtype: string
8
- - name: tgt
9
- dtype: string
10
- - name: bleu
11
- dtype: float64
12
- - name: tgt_index
13
- sequence: int64
14
- - name: original_filename
15
- dtype: string
16
- - name: filename
17
- dtype: string
18
- - name: page_idx
19
- dtype: int64
20
- - name: src_layout
21
- sequence:
22
- sequence: int64
23
- - name: tgt_layout
24
- sequence:
25
- sequence: int64
26
- splits:
27
- - name: dev
28
- num_bytes: 916523238
29
- num_examples: 49994
30
- - name: test
31
- num_bytes: 919351137
32
- num_examples: 50009
33
- - name: train
34
- num_bytes: 7346736959
35
- num_examples: 399997
36
- download_size: 1754136052
37
- dataset_size: 9182611334
38
- - config_name: dev
39
- features:
40
- - name: src
41
- dtype: string
42
- - name: tgt
43
- dtype: string
44
- - name: bleu
45
- dtype: float64
46
- - name: tgt_index
47
- sequence: int64
48
- - name: original_filename
49
- dtype: string
50
- - name: filename
51
- dtype: string
52
- - name: page_idx
53
- dtype: int64
54
- - name: src_layout
55
- sequence:
56
- sequence: int64
57
- - name: tgt_layout
58
- sequence:
59
- sequence: int64
60
- splits:
61
- - name: train
62
- num_bytes: 916523238
63
- num_examples: 49994
64
- - name: dev
65
- num_bytes: 916523238
66
- num_examples: 49994
67
- download_size: 360926516
68
- dataset_size: 1833046476
69
- - config_name: test
70
- features:
71
- - name: src
72
- dtype: string
73
- - name: tgt
74
- dtype: string
75
- - name: bleu
76
- dtype: float64
77
- - name: tgt_index
78
- sequence: int64
79
- - name: original_filename
80
- dtype: string
81
- - name: filename
82
- dtype: string
83
- - name: page_idx
84
- dtype: int64
85
- - name: src_layout
86
- sequence:
87
- sequence: int64
88
- - name: tgt_layout
89
- sequence:
90
- sequence: int64
91
- splits:
92
- - name: test
93
- num_bytes: 919351137
94
- num_examples: 50009
95
- download_size: 181145705
96
- dataset_size: 919351137
97
- configs:
98
- - config_name: default
99
- data_files:
100
- - split: dev
101
- path: data/dev-*
102
- - split: test
103
- path: data/test-*
104
- - split: train
105
- path: data/train-*
106
- - config_name: dev
107
- data_files:
108
- - split: train
109
- path: dev/train-*
110
- - split: dev
111
- path: dev/dev-*
112
- - config_name: test
113
- data_files:
114
- - split: test
115
- path: test/test-*
116
- ---
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ dataset_info:
4
+ - config_name: default
5
+ features:
6
+ - name: src
7
+ dtype: string
8
+ - name: tgt
9
+ dtype: string
10
+ - name: bleu
11
+ dtype: float64
12
+ - name: tgt_index
13
+ sequence: int64
14
+ - name: original_filename
15
+ dtype: string
16
+ - name: filename
17
+ dtype: string
18
+ - name: page_idx
19
+ dtype: int64
20
+ - name: src_layout
21
+ sequence:
22
+ sequence: int64
23
+ - name: tgt_layout
24
+ sequence:
25
+ sequence: int64
26
+ splits:
27
+ - name: dev
28
+ num_bytes: 916523238
29
+ num_examples: 49994
30
+ - name: test
31
+ num_bytes: 919351137
32
+ num_examples: 50009
33
+ - name: train
34
+ num_bytes: 7346736959
35
+ num_examples: 399997
36
+ download_size: 1754136052
37
+ dataset_size: 9182611334
38
+ - config_name: dev
39
+ features:
40
+ - name: src
41
+ dtype: string
42
+ - name: tgt
43
+ dtype: string
44
+ - name: bleu
45
+ dtype: float64
46
+ - name: tgt_index
47
+ sequence: int64
48
+ - name: original_filename
49
+ dtype: string
50
+ - name: filename
51
+ dtype: string
52
+ - name: page_idx
53
+ dtype: int64
54
+ - name: src_layout
55
+ sequence:
56
+ sequence: int64
57
+ - name: tgt_layout
58
+ sequence:
59
+ sequence: int64
60
+ splits:
61
+ - name: train
62
+ num_bytes: 916523238
63
+ num_examples: 49994
64
+ - name: dev
65
+ num_bytes: 916523238
66
+ num_examples: 49994
67
+ download_size: 360926516
68
+ dataset_size: 1833046476
69
+ - config_name: test
70
+ features:
71
+ - name: src
72
+ dtype: string
73
+ - name: tgt
74
+ dtype: string
75
+ - name: bleu
76
+ dtype: float64
77
+ - name: tgt_index
78
+ sequence: int64
79
+ - name: original_filename
80
+ dtype: string
81
+ - name: filename
82
+ dtype: string
83
+ - name: page_idx
84
+ dtype: int64
85
+ - name: src_layout
86
+ sequence:
87
+ sequence: int64
88
+ - name: tgt_layout
89
+ sequence:
90
+ sequence: int64
91
+ splits:
92
+ - name: test
93
+ num_bytes: 919351137
94
+ num_examples: 50009
95
+ download_size: 181145705
96
+ dataset_size: 919351137
97
+ configs:
98
+ - config_name: default
99
+ data_files:
100
+ - split: dev
101
+ path: data/dev-*
102
+ - split: test
103
+ path: data/test-*
104
+ - split: train
105
+ path: data/train-*
106
+ - config_name: dev
107
+ data_files:
108
+ - split: train
109
+ path: dev/train-*
110
+ - split: dev
111
+ path: dev/dev-*
112
+ - config_name: test
113
+ data_files:
114
+ - split: test
115
+ path: test/test-*
116
+ ---
117
+
118
+
119
+ # ReadingBank
120
+
121
+ ReadingBank is a benchmark dataset for reading order detection built with weak supervision from WORD documents, which contains 500K document images with a wide range of document types as well as the corresponding reading order information.
122
+
123
+ Our paper "[LayoutReader: Pre-training of Text and Layout for Reading Order Detection](https://arxiv.org/pdf/2108.11591)" has been accepted by EMNLP 2021.
124
+
125
+ Refer to the official repo for more details: https://github.com/doc-analysis/ReadingBank