ViDeBERTa: A powerful pre-trained language model for Vietnamese

TL;DR

ViDeBERTa is a new Vietnamese language model based on DeBERTa architecture that outperforms previous models on key NLP tasks despite having fewer parameters, advancing Vietnamese NLP capabilities.

Contribution

Introduces ViDeBERTa, a set of pre-trained Vietnamese language models that outperform existing models on multiple NLP tasks with fewer parameters.

Findings

ViDeBERTa surpasses previous state-of-the-art models on Vietnamese NLP tasks.

ViDeBERTa_base achieves comparable or better results with only 23% of PhoBERT_large's parameters.

Models are publicly available for further research and application.

Abstract

This paper presents ViDeBERTa, a new pre-trained monolingual language model for Vietnamese, with three versions - ViDeBERTa_xsmall, ViDeBERTa_base, and ViDeBERTa_large, which are pre-trained on a large-scale corpus of high-quality and diverse Vietnamese texts using DeBERTa architecture. Although many successful pre-trained language models based on Transformer have been widely proposed for the English language, there are still few pre-trained models for Vietnamese, a low-resource language, that perform good results on downstream tasks, especially Question answering. We fine-tune and evaluate our model on three important natural language downstream tasks, Part-of-speech tagging, Named-entity recognition, and Question answering. The empirical results demonstrate that ViDeBERTa with far fewer parameters surpasses the previous state-of-the-art models on multiple Vietnamese-specific natural language understanding tasks. Notably, ViDeBERTa_base with 86M parameters, which is only about 23% of PhoBERT_large with 370M parameters, still performs the same or better results than the previous state-of-the-art model. Our ViDeBERTa models are available at: https://github.com/HySonLab/ViDeBERTa.