Reproduce “LaVy: Vietnamese Multimodal Large Language Model” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chi Tran, Huong Le Thanh
LaVy is a Vietnamese visual assistant, and the authors introduce LaVy-Bench to support comparable evaluation of Vietnamese image-and-language systems. It combines CLIP-Large, a projector, and Vistral 7B, trained with Vietnamese-oriented data. LaVy scores 45.3% on OpenViVQA, versus 27.9% for mBLIP, but TextQA remains difficult and comparisons are limited to the reported tests.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.