Introducing cosmosGPT: Monolingual Training for Turkish Language Models

TL;DR

This paper introduces cosmosGPT, a Turkish language model trained solely on Turkish data, demonstrating promising performance despite its smaller size compared to multilingual models.

Contribution

The study presents cosmosGPT, a monolingual Turkish language model, along with new datasets for fine-tuning and evaluation, and compares its capabilities to existing models.

Findings

cosmosGPT performs well despite being 10 times smaller

New datasets improve Turkish language model evaluation

Monolingual training is a viable alternative to multilingual models

Abstract

The number of open source language models that can produce Turkish is increasing day by day, as in other languages. In order to create the basic versions of such models, the training of multilingual models is usually continued with Turkish corpora. The alternative is to train the model with only Turkish corpora. In this study, we first introduce the cosmosGPT models that we created with this alternative method. Then, we introduce new finetune datasets for basic language models to fulfill user requests and new evaluation datasets for measuring the capabilities of Turkish language models. Finally, a comprehensive comparison of the adapted Turkish language models on different capabilities is presented. The results show that the language models we built with the monolingual corpus have promising performance despite being about 10 times smaller than the others.