Libra: Building Decoupled Vision System on Large Language Models

Yifan Xu, Xiaoshan Yang, Yaguang Song, Changsheng Xu
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:55371-55388, 2024.

Abstract

In this work, we introduce Libra, a prototype model with a decoupled vision system on a large language model (LLM). The decoupled vision system decouples inner-modal modeling and cross-modal interaction, yielding unique visual information modeling and effective cross-modal comprehension. Libra is trained through discrete auto-regressive modeling on both vision and language inputs. Specifically, we incorporate a routed visual expert with a cross-modal bridge module into a pretrained LLM to route the vision and language flows during attention computing to enable different attention patterns in inner-modal modeling and cross-modal interaction scenarios. Experimental results demonstrate that the dedicated design of Libra achieves a strong MLLM baseline that rivals existing works in the image-to-text scenario with merely 50 million training data, providing a new perspective for future multimodal foundation models. Code is available at https://github.com/YifanXu74/Libra.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-xu24ab, title = {Libra: Building Decoupled Vision System on Large Language Models}, author = {Xu, Yifan and Yang, Xiaoshan and Song, Yaguang and Xu, Changsheng}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {55371--55388}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/xu24ab/xu24ab.pdf}, url = {https://proceedings.mlr.press/v235/xu24ab.html}, abstract = {In this work, we introduce Libra, a prototype model with a decoupled vision system on a large language model (LLM). The decoupled vision system decouples inner-modal modeling and cross-modal interaction, yielding unique visual information modeling and effective cross-modal comprehension. Libra is trained through discrete auto-regressive modeling on both vision and language inputs. Specifically, we incorporate a routed visual expert with a cross-modal bridge module into a pretrained LLM to route the vision and language flows during attention computing to enable different attention patterns in inner-modal modeling and cross-modal interaction scenarios. Experimental results demonstrate that the dedicated design of Libra achieves a strong MLLM baseline that rivals existing works in the image-to-text scenario with merely 50 million training data, providing a new perspective for future multimodal foundation models. Code is available at https://github.com/YifanXu74/Libra.} }
Endnote
%0 Conference Paper %T Libra: Building Decoupled Vision System on Large Language Models %A Yifan Xu %A Xiaoshan Yang %A Yaguang Song %A Changsheng Xu %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-xu24ab %I PMLR %P 55371--55388 %U https://proceedings.mlr.press/v235/xu24ab.html %V 235 %X In this work, we introduce Libra, a prototype model with a decoupled vision system on a large language model (LLM). The decoupled vision system decouples inner-modal modeling and cross-modal interaction, yielding unique visual information modeling and effective cross-modal comprehension. Libra is trained through discrete auto-regressive modeling on both vision and language inputs. Specifically, we incorporate a routed visual expert with a cross-modal bridge module into a pretrained LLM to route the vision and language flows during attention computing to enable different attention patterns in inner-modal modeling and cross-modal interaction scenarios. Experimental results demonstrate that the dedicated design of Libra achieves a strong MLLM baseline that rivals existing works in the image-to-text scenario with merely 50 million training data, providing a new perspective for future multimodal foundation models. Code is available at https://github.com/YifanXu74/Libra.
APA
Xu, Y., Yang, X., Song, Y. & Xu, C.. (2024). Libra: Building Decoupled Vision System on Large Language Models. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:55371-55388 Available from https://proceedings.mlr.press/v235/xu24ab.html.

Related Material