Commit
•
1ad65a1
1
Parent(s):
deba89c
Update README.md
Browse files
README.md
CHANGED
@@ -65,11 +65,14 @@ task_ids:
|
|
65 |
|
66 |
### Dataset Summary
|
67 |
|
68 |
-
|
|
|
|
|
|
|
69 |
|
70 |
### Supported Tasks and Leaderboards
|
71 |
|
72 |
-
|
73 |
|
74 |
### Languages
|
75 |
|
@@ -96,15 +99,45 @@ The frequency of the top ten languages in the dataset is shown below:
|
|
96 |
|
97 |
### Data Instances
|
98 |
|
99 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
|
101 |
### Data Fields
|
102 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
103 |
[More Information Needed]
|
104 |
|
105 |
### Data Splits
|
106 |
|
107 |
-
|
108 |
|
109 |
## Dataset Creation
|
110 |
|
|
|
65 |
|
66 |
### Dataset Summary
|
67 |
|
68 |
+
> The digital collections of the SBB contain 153,942 digitized works from the time period of 1470 to 1945.
|
69 |
+
|
70 |
+
> At the time of publication, 28,909 works have been OCR-processed resulting in 4,988,099 full-text pages.
|
71 |
+
For each page with OCR text, the language has been determined by langid (Lui/Baldwin 2012).
|
72 |
|
73 |
### Supported Tasks and Leaderboards
|
74 |
|
75 |
+
This dataset is useful for training language models on historical/OCR'd text.
|
76 |
|
77 |
### Languages
|
78 |
|
|
|
99 |
|
100 |
### Data Instances
|
101 |
|
102 |
+
Each example represents a single page of OCR'd text.
|
103 |
+
|
104 |
+
A single example of the dataset is as follows:
|
105 |
+
|
106 |
+
```python
|
107 |
+
{'file name': '00000045.xml',
|
108 |
+
'language': 'fr',
|
109 |
+
'language_confidence': 0.9999999999910871,
|
110 |
+
'ppn': '646426230',
|
111 |
+
'text': 'Fig. 156 Tirant les sorts au moyen de la divination de Wen-wang',
|
112 |
+
'wc': [0.6125000119,
|
113 |
+
0.4799999893,
|
114 |
+
0.7916666865,
|
115 |
+
0.8066666722,
|
116 |
+
0.7720000148,
|
117 |
+
0.5849999785,
|
118 |
+
0.7580000162,
|
119 |
+
0.9200000167,
|
120 |
+
0.6449999809,
|
121 |
+
0.6060000062,
|
122 |
+
0.6549999714,
|
123 |
+
0.6362500191]}
|
124 |
+
```
|
125 |
+
|
126 |
|
127 |
### Data Fields
|
128 |
|
129 |
+
- 'file name': filename of the original XML file
|
130 |
+
- 'text': OCR'd text for that page of the item
|
131 |
+
- 'wc': the word confidence for each token predicted by the OCR engine
|
132 |
+
- 'ppn': 'Pica production numbers' an internal ID used by the library. See [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.2702544.svg)](https://doi.org/10.5281/zenodo.2702544) for more details.
|
133 |
+
'language': language predicted by `langid.py` (see above for more details)
|
134 |
+
-'language_confidence': confidence score given by `langid.py`
|
135 |
+
|
136 |
[More Information Needed]
|
137 |
|
138 |
### Data Splits
|
139 |
|
140 |
+
This dataset contains only a single split `train`.
|
141 |
|
142 |
## Dataset Creation
|
143 |
|