Comparison of Czech Transformers on Text Classification Tasks

TL;DR

This paper pre-trains and publicly releases Czech-specific Transformer models, compares their performance on text classification tasks, and highlights the limitations of multilingual models for low-resource languages like Czech.

Contribution

It introduces new monolingual Czech Transformers and provides a comparative analysis against existing models for text classification.

Findings

Monolingual Czech Transformers outperform multilingual models in Czech text classification.

Pre-training procedures for Czech Transformers are detailed and reproducible.

Publicly released models facilitate future research in Czech NLP.

Abstract

In this paper, we present our progress in pre-training monolingual Transformers for Czech and contribute to the research community by releasing our models for public. The need for such models emerged from our effort to employ Transformers in our language-specific tasks, but we found the performance of the published multilingual models to be very limited. Since the multilingual models are usually pre-trained from 100+ languages, most of low-resourced languages (including Czech) are under-represented in these models. At the same time, there is a huge amount of monolingual training data available in web archives like Common Crawl. We have pre-trained and publicly released two monolingual Czech Transformers and compared them with relevant public models, trained (at least partially) for Czech. The paper presents the Transformers pre-training procedure as well as a comparison of pre-trained models on text classification task from various domains.