Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1297,4 +1297,56 @@ print('Prop. continued words:', df.cont_prop.mean())
|
|
| 1297 |
We loosely follow the approach of [Rust _et al.](https://arxiv.org/abs/2012.15613) using the fast [UDPipe](https://ufal.mff.cuni.cz/udpipe) to pre-split documents into words and subsequently run the tokenizer over isolated words. For all languages we use the respective November 2023 snapshot from [Wikipedia](wikimedia/wikipedia). Since Wikipedia, by nature, contains significantly more numbers and dates than other text and most tokenizers split those into single digits, we filtered all lone-standing numbers from the documents. Additionally, we removed any documents that still contained non-parsed HTML code (less than 1%).
|
| 1298 |
|
| 1299 |
## Licensing
|
| 1300 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1297 |
We loosely follow the approach of [Rust _et al.](https://arxiv.org/abs/2012.15613) using the fast [UDPipe](https://ufal.mff.cuni.cz/udpipe) to pre-split documents into words and subsequently run the tokenizer over isolated words. For all languages we use the respective November 2023 snapshot from [Wikipedia](wikimedia/wikipedia). Since Wikipedia, by nature, contains significantly more numbers and dates than other text and most tokenizers split those into single digits, we filtered all lone-standing numbers from the documents. Additionally, we removed any documents that still contained non-parsed HTML code (less than 1%).
|
| 1298 |
|
| 1299 |
## Licensing
|
| 1300 |
+
|
| 1301 |
+
We release our curated benchmark and any associated code under [MIT](https://opensource.org/license/mit) license. However, depending on your use case, the licensing conditions of the original [Wikipedia data](https://huggingface.co/datasets/wikimedia/wikipedia#licensing-information) and [UDPipe](https://github.com/ufal/udpipe/tree/udpipe-2?tab=License-1-ov-file) may apply.
|
| 1302 |
+
|
| 1303 |
+
## Supported Languages
|
| 1304 |
+
This dataset currently contains pre-processed data for the following languages:
|
| 1305 |
+
|
| 1306 |
+
| Language | Code |
|
| 1307 |
+
|:-----------|:-------|
|
| 1308 |
+
| Afrikaans | af |
|
| 1309 |
+
| Arabic | ar |
|
| 1310 |
+
| Armenian | hy |
|
| 1311 |
+
| Basque | eu |
|
| 1312 |
+
| Bulgarian | bg |
|
| 1313 |
+
| Catalan | ca |
|
| 1314 |
+
| Croatian | hr |
|
| 1315 |
+
| Czech | cs |
|
| 1316 |
+
| Danish | da |
|
| 1317 |
+
| Dutch | nl |
|
| 1318 |
+
| English | en |
|
| 1319 |
+
| Estonian | et |
|
| 1320 |
+
| Finnish | fi |
|
| 1321 |
+
| French | fr |
|
| 1322 |
+
| German | de |
|
| 1323 |
+
| Greek | el |
|
| 1324 |
+
| Hebrew | he |
|
| 1325 |
+
| Hindi | hi |
|
| 1326 |
+
| Hungarian | hu |
|
| 1327 |
+
| Indonesian | id |
|
| 1328 |
+
| Irish | ga |
|
| 1329 |
+
| Italian | it |
|
| 1330 |
+
| Japanese | ja |
|
| 1331 |
+
| Korean | ko |
|
| 1332 |
+
| Latvian | lv |
|
| 1333 |
+
| Lithuanian | lt |
|
| 1334 |
+
| Marathi | mr |
|
| 1335 |
+
| Norwegian | no |
|
| 1336 |
+
| Persian | fa |
|
| 1337 |
+
| Polish | pl |
|
| 1338 |
+
| Portuguese | pt |
|
| 1339 |
+
| Romanian | ro |
|
| 1340 |
+
| Russian | ru |
|
| 1341 |
+
| Sanskrit | sa |
|
| 1342 |
+
| Serbian | sr |
|
| 1343 |
+
| Slovak | sk |
|
| 1344 |
+
| Slovenian | sl |
|
| 1345 |
+
| Spanish | es |
|
| 1346 |
+
| Swedish | sv |
|
| 1347 |
+
| Tamil | ta |
|
| 1348 |
+
| Telugu | te |
|
| 1349 |
+
| Turkish | tr |
|
| 1350 |
+
| Ukrainian | uk |
|
| 1351 |
+
| Urdu | ur |
|
| 1352 |
+
| Vietnamese | vi |
|