{"id":41077,"date":"2026-05-20T12:27:04","date_gmt":"2026-05-20T10:27:04","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/dataset-indic-hplt-v1-v2-large-scale-multilingual-pretraining-corpora-for-14-indic-languages-english-cc0\/"},"modified":"2026-05-20T12:27:04","modified_gmt":"2026-05-20T10:27:04","slug":"dataset-indic-hplt-v1-v2-large-scale-multilingual-pretraining-corpora-for-14-indic-languages-english-cc0","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/dataset-indic-hplt-v1-v2-large-scale-multilingual-pretraining-corpora-for-14-indic-languages-english-cc0\/","title":{"rendered":"[Dataset] Indic HPLT V1 &amp; V2 \u2014 Large-scale Multilingual Pretraining Corpora For 14 Indic Languages + English (CC0)"},"content":{"rendered":"<p><!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>I&#8217;ve released two large-scale multilingual pretraining datasets on Hugging Face, built from the HPLT v3 high-quality web crawl. Both are <strong>CC0 licensed<\/strong> (public domain) and ready to use with \ud83e\udd17 Datasets.<\/p>\n<h1>\ud83d\udce6 Indic HPLT v1<\/h1>\n<p><strong>~9.8M documents | ~8.4B estimated tokens | 11 languages<\/strong><br \/> \ud83d\udd17 <a href=\"https:\/\/huggingface.co\/datasets\/AM0908\/indic-hplt-v1\">https:\/\/huggingface.co\/datasets\/AM0908\/indic-hplt-v1<\/a><\/p>\n<p>Covers: Hindi, Bengali, Punjabi, Urdu, Tamil, Telugu, Marathi, Gujarati, Malayalam, Kannada, English<\/p>\n<h1>\ud83d\udce6 Indic HPLT v2 (larger successor)<\/h1>\n<p><strong>~34.6M documents | ~25.5B estimated tokens | 14 languages |<\/strong><br \/> \ud83d\udd17 <a href=\"https:\/\/huggingface.co\/datasets\/AM0908\/indic-hplt-v2\">https:\/\/huggingface.co\/datasets\/AM0908\/indic-hplt-v2<\/a><\/p>\n<p>Adds <strong>Nepali, Odia, and Assamese<\/strong> on top of v1, with ~3.5\u00d7 more documents overall.<\/p>\n<h1>\ud83d\udd27 How it was built<\/h1>\n<ul>\n<li>Source: HPLT v3 sorted shards (top-scoring documents by WDS quality score)<\/li>\n<li>Quality filters: 50\u2013100K chars\/doc, max 50% non-alphabetic chars, min avg word length 2.0<\/li>\n<li>Deduplication: SHA-256 exact dedup on all languages + MinHash LSH near-dedup on English (Jaccard \u2265 0.7)<\/li>\n<li>Pipeline code: <a href=\"https:\/\/github.com\/ashtok\/multilingual-hplt-corpus\">https:\/\/github.com\/ashtok\/multilingual-hplt-corpus<\/a><\/li>\n<\/ul><\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/ashtok897\"> \/u\/ashtok897 <\/a> <br \/> <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1tihh62\/dataset_indic_hplt_v1_v2_largescale_multilingual\/\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1tihh62\/dataset_indic_hplt_v1_v2_largescale_multilingual\/\">[comments]<\/a><\/span><\/p><div class='watch-action'><div class='watch-position align-right'><div class='action-like'><a class='lbg-style1 like-41077 jlk' href='javascript:void(0)' data-task='like' data-post_id='41077' data-nonce='72e055e984' rel='nofollow'><img class='wti-pixel' src='https:\/\/www.graviton.at\/letterswaplibrary\/wp-content\/plugins\/wti-like-post\/images\/pixel.gif' title='Like' \/><span class='lc-41077 lc'>0<\/span><\/a><\/div><\/div> <div class='status-41077 status align-right'><\/div><\/div><div class='wti-clear'><\/div>","protected":false},"excerpt":{"rendered":"<p>I&#8217;ve released two large-scale multilingual pretraining datasets on Hugging Face, built from the HPLT v3 high-quality web&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[85],"tags":[],"class_list":["post-41077","post","type-post","status-publish","format-standard","hentry","category-datatards","wpcat-85-id"],"_links":{"self":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41077","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/comments?post=41077"}],"version-history":[{"count":0,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41077\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/media?parent=41077"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/categories?post=41077"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/tags?post=41077"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}