BamiBERT: A New BERT-based Language Model for Vietnamese
For Vietnamese NLP researchers and practitioners, BamiBERT provides a stronger and more convenient pre-trained encoder, though it is an incremental improvement over existing methods.
BamiBERT is a new BERT-based language model for Vietnamese that outperforms PhoBERT across multiple benchmarks, achieving state-of-the-art results on 11 of 15 metrics while supporting longer contexts and eliminating the need for word segmentation.
In this paper, we introduce BamiBERT, a new BERT-based pre-trained language model for Vietnamese that addresses key limitations of PhoBERT -- the current de facto Vietnamese text encoder. Trained from scratch on a 129GB corpus of general-domain Vietnamese text for 20 epochs, BamiBERT supports an extended context length of up to 2048 tokens and operates directly on raw input, eliminating the need for external word segmentation. Across 8 Vietnamese benchmarks, it achieves the best score on 11 of 15 metrics and the second-best on 3 others, setting a new state of the art among "base"-sized Vietnamese encoders and demonstrating strong cross-domain generalization. We release BamiBERT at: https://huggingface.co/Qualcomm-AI-Research/BamiBERT