Search papers, labs, and topics across Lattice.
M-MiniGPT4, a multilingual vision-language model, was created by training a MiniGPT4 architecture on a mixture of native multilingual and translated data. A novel multilingual alignment training stage using parallel text corpora further enhances the model's cross-lingual capabilities. The resulting model achieves 36% accuracy on the multilingual MMMU benchmark, surpassing other models in its weight class.
Multilingual vision-language models can achieve surprisingly strong performance (36% on MMMU) simply by training on translated data and aligning with parallel text corpora.
This paper presents a Multilingual Vision Large Language Model, named M-MiniGPT4. Our model exhibits strong vision-language understanding (VLU) capabilities across 11 languages. We utilize a mixture of native multilingual and translated data to push the multilingual VLU performance of the MiniGPT4 architecture. In addition, we propose a multilingual alignment training stage that uses parallel text corpora to further enhance the multilingual capabilities of our model. M-MiniGPT4 achieves 36% accuracy on the multilingual MMMU benchmark, outperforming state-of-the-art models in the same weight class, including foundation models released after the majority of this work was completed. We open-source our models, code, and translated datasets to facilitate future research in low-resource and multilingual settings.
