Code & Resources

All our research code is available on GitHub.

LinguaFlow

Neural Architecture for Multilingual NLP

A comprehensive toolkit for building and evaluating multilingual language models with cross-lingual transfer capabilities.

DialogueMaster

End-to-End Dialogue Systems

Open-source framework for building task-oriented and open-domain conversational AI systems.

KnowledgeGraph-QA

Question Answering over Knowledge Graphs

State-of-the-art models for semantic parsing and question answering over large-scale knowledge bases.

FairNLP

Fairness Evaluation Toolkit

Tools and benchmarks for evaluating and improving fairness in NLP models.

Datasets

Dataset Description Size Download
MultiLang-QA Multilingual QA benchmark 500K examples Link
DialogueBench Dialogue evaluation dataset 100K conversations Link
FairText Bias evaluation corpus 50K sentences Link

Pre-trained Models

We release pre-trained models on Hugging Face:

Tutorials