Datasets:
update readme
Browse files
README.md
CHANGED
@@ -91,7 +91,7 @@ purpose approaches.
|
|
91 |
|
92 |
### Supported Tasks and Leaderboards
|
93 |
|
94 |
-
|
95 |
|
96 |
### Languages
|
97 |
|
@@ -101,15 +101,29 @@ Basque (eu)
|
|
101 |
|
102 |
### Data Instances
|
103 |
|
104 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
105 |
|
106 |
### Data Fields
|
107 |
|
108 |
-
|
|
|
|
|
|
|
|
|
|
|
109 |
|
110 |
### Data Splits
|
111 |
|
112 |
-
|
113 |
|
114 |
## Dataset Creation
|
115 |
|
@@ -174,6 +188,7 @@ a.soroa@ehu.eus.
|
|
174 |
### Citation Information
|
175 |
|
176 |
If you use our corpus or models for academic research, please cite the paper in question:
|
|
|
177 |
@misc{artetxe2022euscrawl,
|
178 |
title={Does corpus quality really matter for low-resource languages?},
|
179 |
author={Mikel Artetxe, Itziar Aldabe, Rodrigo Agerri,
|
@@ -183,6 +198,7 @@ If you use our corpus or models for academic research, please cite the paper in
|
|
183 |
archivePrefix={arXiv},
|
184 |
primaryClass={cs.CL}
|
185 |
}
|
|
|
186 |
|
187 |
### Contributions
|
188 |
|
|
|
91 |
|
92 |
### Supported Tasks and Leaderboards
|
93 |
|
94 |
+
EusCrawl is intended for pretraining models for language modeling or masked language modeling.
|
95 |
|
96 |
### Languages
|
97 |
|
|
|
101 |
|
102 |
### Data Instances
|
103 |
|
104 |
+
```json
|
105 |
+
{
|
106 |
+
"id": 6,
|
107 |
+
"title": "Herriko enpresa handien eta txikien arteko topaketak egingo dituzte",
|
108 |
+
"text": "09:30ean hasiko da bilera eta aurkezpena egingo dute Tubacex, JEZ, Envases, Guardian eta Vidrala enpresek. Eskualdeko lantegi motorrekin beste enpresa txikiak eta ertainak egongo dira. Erakunde publikoaren helburua da euren artean ezagutzea eta elkarlana sustatzea.",
|
109 |
+
"source": "aiaraldea",
|
110 |
+
"license": "cc-by-sa 3.0",
|
111 |
+
"url": "https://aiaraldea.eus/laudio/1494603159768-herriko-enpresa-handien-eta-txikien-arteko-topaketak-egingo-dituzte",
|
112 |
+
}
|
113 |
+
```
|
114 |
|
115 |
### Data Fields
|
116 |
|
117 |
+
- "id": example id
|
118 |
+
- "title": article title
|
119 |
+
- "text": article text
|
120 |
+
- "source": article source
|
121 |
+
- "license": article license
|
122 |
+
- "url": article url
|
123 |
|
124 |
### Data Splits
|
125 |
|
126 |
+
The dataset only has one training split because it is intended for pretraining language models.
|
127 |
|
128 |
## Dataset Creation
|
129 |
|
|
|
188 |
### Citation Information
|
189 |
|
190 |
If you use our corpus or models for academic research, please cite the paper in question:
|
191 |
+
```bibtex
|
192 |
@misc{artetxe2022euscrawl,
|
193 |
title={Does corpus quality really matter for low-resource languages?},
|
194 |
author={Mikel Artetxe, Itziar Aldabe, Rodrigo Agerri,
|
|
|
198 |
archivePrefix={arXiv},
|
199 |
primaryClass={cs.CL}
|
200 |
}
|
201 |
+
```
|
202 |
|
203 |
### Contributions
|
204 |
|