Part of International Conference on Representation Learning 2025 (ICLR 2025) Conference
Feng Li, Renrui Zhang, Hao Zhang, Yuanhan Zhang, Bo Li, Wei Li, Zejun MA, Chunyuan Li
Visual instruction tuning has made considerable strides in enhancing the capabilities of Large Multimodal Models (LMMs). However, existing open LMMs largely focus on single-image tasks, their applications to multi-image scenarios remains less explored. Additionally, prior LMM research separately tackles different scenarios, leaving it impossible to generalize cross scenarios with newemerging capabilities. To this end, we introduce LLaVA-Interleave, which simultaneously tackles Multi-image, Multi-frame (video), Multi-view (3D), and Multi-patch (single-image) scenarios in LMMs. To enable these capabilities, we regard the interleaved data format as a general template and compile the M4-Instruct dataset with 1,177.6k samples, spanning 4 primary domains with 14tasks and 41 datasets. We also curate the LLaVA-Interleave Bench to comprehensively evaluate the multi-image performance of LMMs. Through extensiveexperiments, LLaVA-Interleave achieves leading results in multi-image, video,and 3D benchmarks, while maintaining the performance of single-image tasks.Besides, our model also exhibits several emerging capabilities, e.g., transferring tasks across different settings and modalities.