low-resource-language

🇧🇮 The first large-scale, open-source speech and text dataset for Kirundi language. Building AI models for 12M+ Kirundi speakers through community collaboration. Includes ASR, TTS, and MT capabilities.

⌨️ Klavyé Kréyòl Karukera — clavier Android intelligent pour le créole guadeloupéen : suggestions bilingues kréyòl/français, accents, 100% hors ligne. Guadeloupean Creole keyboard.

Open-source Windows dictation and AI-assisted text correction for Georgian, a low-resource language.

Multilingual LLM safety benchmark for Urdu, Pashto, and code-switching prompts across different prompting techniques

General-purpose SigerLM from scratch with SSM architecture, LoRA, long-context training, frontend vision, design-to-code, and Lampung language support.

UrduReason-Eval: A comprehensive evaluation dataset with 800 Urdu reasoning problems across 6 categories (arithmetic, logical deduction, temporal, comparative, and causal reasoning) for assessing reasoning capabilities in Urdu language models.

Telugu AI Voice & Text Summarizer 2026 🔥 mT5 & TTS NLP App

Com la tokenització fractura la morfologia catalana i si una segmentació conscient dels morfemes recupera la geometria. Provat en 3 llengües indoeuropees (català, castellà, anglès): el català es fragmenta ~1,7× més que l'anglès; forçar el tall morfèmic recupera la composicionalitat (robust a portadora i replicat en castellà).

Traduction, reconnaissance vocale (ASR) et synthèse vocale (TTS) pour l'éwé, langue peu dotée du Togo/Ghana : pipeline complet de collecte, nettoyage, fine-tuning (NLLB-200 + LoRA, MMS), évaluation (BLEU 26 ewe→en, WER 29 %) et démo Streamlit. Modèles publiés sur Hugging Face.

LLM-Based Evaluation of Low-Resource Machine Translation (CLNLP 2025)

A repository dedicated to compiling Jaqaru texts and build a Spanish-Jaqaru parallel corpus.

AI-agent skills and sub-agents for extracting quality-checked parallel corpora from local-language dictionary PDFs — language-agnostic OCR pipeline with typo correction, meaning cross-checking, and multi-book merging

FilBench website: filbench.github.io

This project will explore how well existing ASR systems perform for African languages ( Setswana ) and how extracted transcripts from local language podcasts can be used for topic modeling and how what NLP techniques improve topic modeling.

Final version of Bartangi Lemmatizer and Word2Vec embeddings