Corpora & data

Text and speech corpora, parallel datasets and language archives for research and language technology.

10entriesSource review current through 17 Jul 2026.

Search the catalog

Filters

Showing 10 of 10 entries

buryat-translation/buryat_monocorpus (Hugging Face)

A downloadable monolingual Buryat corpus of 215,000 lines of text (May 2025) — a raw database for language models and lexicographical work on Buryat.

DatasetOpen access

Corpora & dataLive

buryat-translation/buryat_russian_parallel_corpus (Hugging Face)

A downloadable parallel Buryat-Russian corpus of 120,000 sentence pairs (May 2025); its size matches the corpus transferred to the Ministry of Education and Science of Buryatia / Belig Centre (Yandex) in April 2025 for the implementation of Buryat into Yandex Translate.

DatasetOpen access

Corpora & dataLive

ELAR / ELDP (Endangered Languages Archive)

An archive for the documentation of endangered languages (under the Berlin-Brandenburg Academy of Sciences and Humanities); while there is no large, separate Buryat deposit yet, the ELDP platform and grant programme represent a real pathway to addressing the gap in dialect documentation (Sartul, Songol, Shenehen).

ArchiveOpen access

Corpora & dataLive

Electronic Corpus of the Buryat Language (IMBT SB RAS)

The largest written corpus of literary Buryat (~3 million word occurrences, automorph analysis ~85%, lemmatisation), maintained by IMBT SB RAS and updated since 2012 (last update: October 2024); designed as the foundation for a future National Corpus.

CorpusOpen access

Corpora & dataLive

kibaraki/Shinekhen-Buryat (Hugging Face)

The only identified open dataset for the Shenehen dialect (1,250 lines, September 2025)—addresses a gap that the main landscape report noted as entirely absent regarding Buryat dialects.

DatasetOpen access

Corpora & dataLive

lingtrain/buryat-russian (Hugging Face)

A small downloadable parallel Buryat-Russian corpus (1,330 lines)—another open dataset for MT/NLP experiments with the Buryat language.

DatasetOpen access

Corpora & dataPartial access

NKFI-EPR — registry of a grant for an etymological dictionary of the Buryat language

An entry in the Hungarian state grant registry regarding the funding of a project for an etymological dictionary of the Buryat language.

DatasetOpen access

Corpora & dataLive

SaranaAbidueva/buryat_monocorpus (Hugging Face)

A monolingual Buryat corpus (220,000 lines) from the same project by Sarana Abiduyeva as the paired Russian-Buryat dataset.

DatasetOpen access

Corpora & dataLive

SaranaAbidueva/buryat-russian_parallel_corpus (Hugging Face)

A parallel Buryat-Russian corpus (41,000 pairs) uploaded by Sarana Abiduyeva—the author of a neural network Russian-Buryat translator created with grant support from the Government of Buryatia; data from this project formed the basis for the appearance of the Buryat language in Yandex Translate (from 2025, source: belig.ru).

DatasetOpen access

Corpora & dataLive

tsakorpus (IMBT corpus engine)

An open-source engine based on Elasticsearch upon which the IMBT Buryat language electronic corpus is built; useful as a technical reference for those wishing to understand or reuse the corpus infrastructure.

AppOpen access