The definitive index of Bangla NLP research
Papers, datasets, models, tools, and benchmarks for Bengali natural language processing — community-maintained, link-verified, and open on GitHub.
Browse by task · কাজ অনুযায়ী
All tasks →Recently added
most recently verified · mainPowered by the community · সম্প্রদায়
Bangla is low-resource, and this catalog exists because a handful of groups, labs, and maintainers release their work openly. These are some whose datasets, models, and tools the Hub is built on — credited by real contribution, not ranking.
The research group behind much of the modern Bangla NLP stack.
Community non-profit running open speech and handwriting datasets and competitions.
Builds and openly releases Bangla large language models.
Releases code-mixed and offensive-language datasets for Bangla.
Maintainer of the BNLP toolkit and early open Bangla models.
Co-authored the Bengali physics MCQ solver using LLM chain-of-thought reasoning.
Co-authored the Bengali physics MCQ solver using LLM chain-of-thought reasoning.
Co-authored the Bengali physics MCQ solver using LLM chain-of-thought reasoning.
Co-authored the Bengali physics MCQ solver using LLM chain-of-thought reasoning.
Released the Lipi-Ghor long-form, multi-speaker Bengali speech corpus (ASR + diarization).
Missing someone whose open work belongs here? Add them via PR — with a concrete contribution, not a claim.
Built by · নির্মাতা
The people who build and maintain the Hub itself — the catalog code, ingestion tooling, and views. Credited from the repo's git history, by merged contribution.
Creator and maintainer — built the Astro catalog, the ingestion pipeline, and every view.
Refactored the contributor, model, and leaderboard data layers, tightened link checking, fixed catalog consistency and mobile navigation, audited dataset counts, citations, and licenses, and added keyboard search and shareable catalog filters.
Added the Speech Emotion Recognition and Speaker tasks, and fixed the Home hero search dropdown layering.