@RIOTU_LAB
About Our Project
The ArabianLLM project at Prince Sultan University's Robotics and Internet-of-Things Lab is an innovative effort in Arabic Language Modeling. It aims to develop specialized Large Language Models for Arabic, tackling its distinctive morphological and script challenges.
Expertly crafted for enhanced efficiency and accuracy in Arabic language modeling.
Focused on innovative solutions for Arabic text generation and modeling.
Our team brings a fresh perspective and dedicated approach to the field of computational linguistics and AI, making significant strides in Arabic NLP and Language Modeling.
Explore our cutting-edge projects in Arabic NLP, showcasing specialized models and tokenization tools designed for deep linguistic analysis.
ArabianGPT 0.1B, part of the ArabianLLM series, with 134M parameters, 12 layers, and 768 token context window, trained on the Abu Elkhiar Corpus for news content processing.
ArabianGPT 0.3B features 345M parameters, 24 layers, 16 MALs, and a 1024 token context window, adept at capturing complex Arabic nuances across various domains.
Aranizer, with SentencePiece and BPE, enhances Arabic NLP with variants up to 86K vocabulary, significantly improving precision and F1 scores in text analysis.
Meet the dedicated professionals behind our groundbreaking work in NLP and AI at the RIOTU Lab, Prince Sultan University.
1
2
3
4
5
Discover the cutting-edge developments in Arabic NLP with our latest publications and technical reports from the RIOTU Lab team.
Aranizer Tokenizer
Aranizer: Leveraging SentencePiece and Byte Pair Encoding for Arabic Text Tokenization
ArabianGPT
ArabianLLM: Native Arabic Large Language Modeling for Text Generation
© Robotics and Internet-of-Things Lab.
