Datasets:

procesaur
/

kisobran

	---
	license: cc-by-sa-4.0
	task_categories:
	- text-generation
	language:
	- sr
	- hr
	- bs
	tags:
	- webdataset
	pretty_name: Kišobran (Umbrella corp.)
	size_categories:
	- 10B<n<100B
	configs:
	- config_name: default
	data_files:
	- split: train
	path: '*.txt'
	- split: sr
	path: '*_sr.txt'
	- split: cnr
	path: '*_cnr.txt'
	- split: hr
	path: '*_hr.txt'
	- split: bs
	path: '*_bs.txt'
	---

	<img src="cover.png" class="cover">


	<table style="width:100%;height:100%">
	<!--tr style="width:100%;height:30px">
	<td colspan=2 align=center>
	<h1>Kišobran (Umbrella corp.)</h1>
	</td>
	<tr-->
	<tr style="width:100%;height:100%">
	<td width=50%>
	<h2><span class="highlight-container"><b class="highlight">Kišobran korpus</b></span> - krovni veb korpus srpskog i srpskohrvatskog jezika</h2>
	<p>Najveća agregacija veb korpusa do sada, pogodna za obučavanje velikih jezičkih modela za srpski jezik.</p>
	<p>Ukupno x dokumenata, ukupno sa <span class="highlight-container"><span class="highlight">preko 18.5 milijardi reči</span></span>.</p>
	<p></p>
	<p>Svaka linija predstavlja novi dokument</p>
	<p>Rečenice unutar dokumenata su obeležene.</p>
	<h4>Sadrži obrađene i deduplikovane verzije sledećih korpusa:</h4>
	</td>
	<td>
	<h2><span class="highlight-container"><b class="highlight">Umbrella corp.</b></span> - umbrella web corpus of Serbian and Serbo-Croatian</h2>
	<p>The largest aggregation of web corpora so far, suitable for training Serbian large language models.</p>
	<p>A total of x documents containing <span class="highlight-container"><span class="highlight">over 18.5 billion words</span></span>.</p>
	<p></p>
	<p>Each line represents a document.</p>
	<p>Each Sentence in a document is delimited.</p>
	<h4>Contains processed and deduplicated versions of the following corpora:</h4>
	</td>
	</tr>
	</table>


	<table class="lista">
	<tr>
	<td>Korpus<br/>Corpus</td>
	<td>Jezik<br/>Language</td>
	<td>Broj reči<br/>Word count</td>
	<td>Broj dokumenata<br/>Doc. count</td>
	<td>Udeo<br/>Share</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/HPLT/hplt_monolingual_v1_2">HPLT_sr</a></td>
	<td>🇷🇸</td>
	<td>2.9 M</td>
	<td>2.5 B</td>
	<td>13.74%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1807">MaCoCu_sr</a></td>
	<td>🇷🇸</td>
	<td>6.7 M</td>
	<td>2.1 B</td>
	<td>11.54%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/allenai/c4">MC4_sr</a></td>
	<td>🇷🇸</td>
	<td>2.3 M</td>
	<td>782 M</td>
	<td>4.19%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/cc100">cc100_sr</a></td>
	<td>🇷🇸</td>
	<td>2.3 M</td>
	<td>659 M</td>
	<td>3.53%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1752">PDRS1.0</a></td>
	<td>🇷🇸</td>
	<td>400 K</td>
	<td>506 M</td>
	<td>2.71%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/jerteh/SrpKorNews">SrpKorNews</a></td>
	<td>🇷🇸</td>
	<td>35 K</td>
	<td>469 M</td>
	<td>2.51%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/oscar-corpus/OSCAR-2301">OSCAR_sr</a></td>
	<td>🇷🇸</td>
	<td>500 K</td>
	<td>410 M</td>
	<td>2.2%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1063">srWaC</a></td>
	<td>🇷🇸</td>
	<td>1.2 M</td>
	<td>307 M</td>
	<td>1.65%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1426">CLASSLA_sr</a></td>
	<td>🇷🇸</td>
	<td>1.3 M</td>
	<td>240 M</td>
	<td>1.29%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1809">MaCoCu_cnr</a></td>
	<td>🇷🇸/🇲🇪</td>
	<td>500 K</td>
	<td>152 M</td>
	<td>0.82%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1429">meWaC</a></td>
	<td>🇷🇸/🇲🇪</td>
	<td>200 K</td>
	<td>41 M</td>
	<td>0.22%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/cc100">cc100_hr</a></td>
	<td>🇭🇷</td>
	<td>13.3 M</td>
	<td>2.5 B</td>
	<td>13.73%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1806">MaCoCu_hr</a></td>
	<td>🇭🇷</td>
	<td>8 M</td>
	<td>2.3 B</td>
	<td>12.63%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/HPLT/hplt_monolingual_v1_2">HPLT_hr</a></td>
	<td>🇭🇷</td>
	<td>2.3 M</td>
	<td>1.8 B</td>
	<td>9.95%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/classla/xlm-r-bertic-data">hr_news</a></td>
	<td>🇭🇷</td>
	<td>4.1 M</td>
	<td>1.4 B</td>
	<td>7.65%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1064">hrWaC</a></td>
	<td>🇭🇷</td>
	<td>3.1 M</td>
	<td>935 M</td>
	<td>5.01%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1426">CLASSLA_hr</a></td>
	<td>🇭🇷</td>
	<td>1.2 M</td>
	<td>160 M</td>
	<td>0.86%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1180">riznica</a></td>
	<td>🇭🇷</td>
	<td>20 K</td>
	<td>69 M</td>
	<td>0.37%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1808">MaCoCu_bs</a></td>
	<td>🇧🇦</td>
	<td>2.6 M</td>
	<td>700 M</td>
	<td>3.75%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1062">bsWaC</a></td>
	<td>🇧🇦</td>
	<td>800 K</td>
	<td>194 M</td>
	<td>1.04%</td>
	</tr>
	<tr>
	<td><a href="https://www.clarin.si/repository/xmlui/handle/11356/1426">CLASSLA_bs</a></td>
	<td>🇧🇦</td>
	<td>800 K</td>
	<td>105 M</td>
	<td>0.56%</td>
	</tr>
	<tr>
	<td><a href="https://huggingface.co/datasets/cc100">cc100_bs</a></td>
	<td>🇧🇦</td>
	<td>300 K</td>
	<td>9 M</td>
	<td>0.05%</td>
	</tr>
	<tr>
	<td><b>TOTAL</b></td>
	<td></td>
	<td><b>54.75 M</b></td>
	<td><b>18.65 B</b></td>
	<td>100%</td>
	</tr>
	</table>

	Load complete dataset / Učitavanje kopletnog dataseta
	```python
	from datasets import load_dataset
	dataset = load_dataset("procesaur/umbrella")
	```


	Load a specific language / Učitavanje pojedinačnih jezika
	```python
	from datasets import load_dataset
	dataset_sr = load_dataset("procesaur/umbrella", "sr")
	dataset_cnr = load_dataset("procesaur/umbrella", "cnr")
	dataset_hr = load_dataset("procesaur/umbrella", "hr")
	dataset_bs = load_dataset("procesaur/umbrella", "bs")
	```


	<div class="inline-flex flex-col" style="line-height: 1.5;padding-right:50px">
	<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">Editor</div>
	<a href="https://huggingface.co/procesaur">
	<div class="flex">
	<div
	style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%;
	background-size: cover; background-image: url('https://cdn-uploads.huggingface.co/production/uploads/1673534533167-63bc254fb8c61b8aa496a39b.jpeg?w=200&h=200&f=face')">
	</div>
	</div>
	</a>
	<div style="text-align: center; font-size: 16px; font-weight: 800">Mihailo Škorić</div>
	<div>
	<a href="https://huggingface.co/procesaur">
	<div style="text-align: center; font-size: 14px;">@procesaur</div>
	</a>
	</div>
	</div>
	</div>



	Citation:
	```bibtex
	@article{skoric24korpusi,
	author = {\vSkori\'c, Mihailo and Jankovi\'c, Nikola},
	title = {New Textual Corpora for Serbian Language Modeling},
	journal = {Infotheca},
	volume = {24},
	issue = {1},
	year = {2024},
	publisher = {Zajednica biblioteka univerziteta u Srbiji, Beograd}
	}
	```


	<table style="width:100%;height:100%">
	<tr style="width:100%;height:100%">
	<td width=50%>
	<p>Istraživanje je sprovedeno uz podršku Fonda za nauku Republike Srbije, #7276, Text Embeddings – Serbian Language Applications – TESLA.</p>
	<p>Svaki korpus u tabeli vezan je za URL sa kojeg je preuzet. Prikazani brojevi dokumenata i reči, odnose se na stanje nakon čićenja i deduplikacije.</p>
	<p>Deduplikacija je izvršena pomoću alata <a href="http://corpus.tools/wiki/Onion">onion</a> korišćenjem pretrage 6-torki i pragom dedumplikacije 75%.</p>
	<p>Računarske resursre neophodne za deduplikaciju korpusa obezbedila je Nacionalna platforma za veštačku inteligenciju Srbije.</p>
	</td>
	<td>
	<p>This research was supported by the Science Fund of the Republic of Serbia, #7276, Text Embeddings - Serbian Language Applications - TESLA.</p>
	<p>Each corpus in the table is linked to the URL from which it was downloaded. The displayed numbers of documents and words refer to after cleaning and deduplication.</p>
	<p>The dataset was deduplicated using <a href="http://corpus.tools/wiki/Onion">onion</a> using 6-tuples search and a duplicate threshold of 75%.</p>
	<p>Computer resources necessary for the deduplication of the corpus were provided by the National Platform for Artificial Intelligence of Serbia.</p>
	</td>
	</tr>
	</table>



	<div id="zastava">
	<div class="grb">
	<img src="https://www.ai.gov.rs/img/logo_60x120-2.png" style="position:relative; left:30px; z-index:10; height:85px">
	</div>
	<table width=100% style="border:0px">
	<tr style="background-color:#C6363C;width:100%;border:0px;height:30px"><td></td></tr>
	<tr style="background-color:#0C4076;width:100%;border:0px;height:30px"><td></td></tr>
	<tr style="background-color:#ffffff;width:100%;border:0px;height:30px"><td></td></tr>
	</table>
	</div>

	<style>
	.ffeat: {
	color:red
	}

	.cover {
	width: 100%;
	margin-bottom: 5pt
	}

	.highlight-container, .highlight {
	position: relative;
	text-decoration:none
	}

	.highlight-container {
	display: inline-block;

	}

	.highlight{
	color:white;
	text-transform:uppercase;
	font-size: 16pt;
	}

	.highlight-container{
	padding:5px 10px
	}

	.highlight-container:before {
	content: " ";
	display: block;
	height: 100%;
	width: 100%;
	margin-left: 0px;
	margin-right: 0px;
	position: absolute;
	background: #e80909;
	transform: rotate(2deg);
	top: -1px;
	left: -1px;
	border-radius: 20% 25% 20% 24%;
	padding: 10px 18px 18px 10px;
	}

	div.grb, #zastava>table {
	position:absolute;
	top:0px;
	left: 0px;
	margin:0px
	}

	div.grb>img, #zastava>table{
	margin:0px
	}

	#zastava {
	position: relative;
	margin-bottom:120px
	}

	p {
	font-size:14pt
	}
	.lista tr{
	line-height:1
	}
	</style>