Generative large language models (LLMs) exhibit impressive capabilities, which can be further augmented by integrating a pre-trained vision model into the original LLM to create a multimodal LLM (MLLM). However, this integration often significantly decreases performance on natural language understanding and generation tasks, compared to the original LLM. This study investigates this issue using the LLaVA MLLM, treating the integration as a continual learning problem. We evaluate five continual learning methods to mitigate forgetting and identify a technique that enhances visual understanding while minimizing linguistic performance loss. Our approach reduces linguistic performance degradation by up to 15% over the LLaVA recipe, while maintaining high multimodal accuracy. We also demonstrate the robustness of our method through continual learning on a sequence of vision-language tasks, effectively preserving linguistic skills while acquiring new multimodal capabilities. Project webpage: https://shikhar-srivastava.github.io/cl-for-improving-mllms
@article{arxiv.2410.19925,
title = {Improving Multimodal Large Language Models Using Continual Learning},
author = {Shikhar Srivastava and Md Yousuf Harun and Robik Shrestha and Christopher Kanan},
journal= {arXiv preprint arXiv:2410.19925},
year = {2025}
}
Comments
CoLLAs 2025 and Scalable Continual Learning for Lifelong Foundation Models, NeurIPS 2024