Teuta by LTS-VVE is a 3.2 billion parameter bilingual instruction-tuned language model based on Llama-3.2-3B, designed for question answering in Albanian and English. It is fine-tuned on diverse datasets covering mathematics, philosophy, chemistry, biology, code (especially Rust), psychology, and climate science. This model excels at multilingual applications and under-resourced language support, with a strong focus on Albanian. It is particularly suited for research, educational tools, and domain-specific applications requiring instruction-following capabilities.
No reviews yet. Be the first to review!