paper-with-me

Papers

Enhancing Multimodal Continual Instruction Tuning with BranchLoRA

2025-05-31 · Duzhen Zhang, Yong Ren, Zhong-Zhi Li, Yahan Yu, Jiahua Dong, Chenxing Li, Zhilong Ji, Jinfeng Bai

Multimodal Continual Instruction Tuning (MCIT) aims to finetune Multimodal Large Language Models (MLLMs) to continually align with human intent across sequential tasks. Existing approaches often rely on the Mixture-of-Experts (MoE) LoRA framework to preserve previous instruction alignments. However, these methods are prone to Catastrophic Forgetting (CF), as they aggregate all LoRA blocks via simple summation, which compromises performance over time. In this paper, we identify a critical parameter inefficiency in the MoELoRA framework within the MCIT context. Based on this insight, we propose BranchLoRA, an asymmetric framework to enhance both efficiency and performance. To mitigate CF, we introduce a flexible tuning-freezing mechanism within BranchLoRA, enabling branches to specialize in intra-task knowledge while fostering inter-task collaboration. Moreover, we incrementally incorporate task-specific routers to ensure an optimal branch distribution over time, rather than favoring the most recent task. To streamline inference, we introduce a task selector that automatically routes test inputs to the appropriate router without requiring task identity. Extensive experiments on the latest MCIT benchmark demonstrate that BranchLoRA significantly outperforms MoELoRA and maintains its superiority across various MLLM sizes.

📄 PDF Abstract BibTeX arXiv:2506.02041

Code (0)

등록된 구현이 없습니다.

Tasks

Mixture-of-Experts

Methods 이 논문이 사용한 방법론

ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

Continual Instruction Tuning for Large Multimodal Models

2023-11-27 · Jinghan He, Haiyun Guo, Ming Tang, Jinqiao Wang

Instruction tuning is now a widely adopted approach to aligning large multimodal models (LMMs) to follow human intent. It unifies the data format of vision-language tasks, enabling multi-task joint training. However, vis…

Continual Learning

MCITlib: Multimodal Continual Instruction Tuning Library and Benchmark

2025-08-10 · Haiyang Guo, Fei Zhu, Hongbo Zhao, Fanhu Zeng 외 arxiv

Continual learning enables AI systems to acquire new knowledge while retaining previously learned information. While traditional unimodal methods have made progress, the rise of Multimodal Large Language Models (MLLMs) b…

Continual Learning

Multimodal Continual Instruction Tuning with Dynamic Gradient Guidance

2025-11-19 · Songze Li, Mingyu Gao, Tonghua Su, Xu-Yao Zhang 외 arxiv

Multimodal continual instruction tuning enables multimodal large language models to sequentially adapt to new tasks while building upon previously acquired knowledge. However, this continual learning paradigm faces the s…

Continual Learning

SwitchCIT: Switching for Continual Instruction Tuning

2024-07-16 · Xinbo Wu, Max Hartman, Vidhata Arjun Jayaraman, Lav R. Varshney

Large language models (LLMs) and multimodal models (MMs) have exhibited impressive capabilities in various domains, particularly in general language understanding and visual reasoning. However, these models, trained on m…

Text GenerationVisual Reasoning

LLaVA-c: Continual Improved Visual Instruction Tuning

2025-06-10 · Wenzhuo LIU, Fei Zhu, Haiyang Guo, Longhui Wei 외

Multimodal models like LLaVA-1.5 achieve state-of-the-art visual understanding through visual instruction tuning on multitask datasets, enabling strong instruction-following and multimodal performance. However, multitask…

Continual LearningContinual PretrainingInstruction Following