CoChat

MoE-LLaVA : Mixture of Experts for Large Vision-Language Models

Open access
Lin, Bin, Tang, Zhenyu, Ye, Yang, Huang, Jinfa, Zhang, Junwu, Pang, Yatian, Jin, Peng, Ning, Munan, Luo, Jiebo, Yuan, Li · 2026 · IEEE Transactions on Multimedia
Cited by 69

The DOI matches its Crossref record; 20 of the 20 references it cites were found in a registry or an index.

Abstract

Recently, remarkable progress has been made in scaling up Large Language Models (LLMs) through the use of the sparse Mixture-of-Expert (MoE) layers without significantly increasing computational cost. However, the transition from a pre-trained LLM to a sparse Large Vision-Language Model (LVLM) with MoE remains an open challenge. Directly fine-tuning an LLM to a sparse LVLM often leads to training collapse, characterized by (1) a large modality feature distribution gap and (2) expert load imbalance. This paper proposes a three-stage decoupled weight training process. In the first two stages, the model learns to adapt the LLM to an LVLM. In the third stage, the FFN weights from the second stage are used as lossless initialization for expert weights, effectively constructing a sparse model with a vast number of parameters while maintaining constant computational cost. Through extensive ablation experiments, we derive three empirical guidelines and propose a sparse LVLM termedMoE-LLaVA. MoE-LLaVA is a MoE-based sparse LVLM architecture, which uniquely activates only the top-$k$experts through routers during deployment, keeping the remaining experts inactive. Extensive experiments demonstrate that MoE-LLaVA outperforms LLaVA-1.5-7B with an average improvement of 4.6 across nine visual understanding benchmarks. Notably, with only 2.2B active parameters, our MoE-LLaVA shows comparable result with LLaVA-1.5-13B (87.0 vs. 85.9) on POPE benchmark. Our work establishes a baseline for sparse LVLMs and provides empirical guidelines for exploring the sparse LVLMs. Our code is available at:https://github.com/PKU-YuanGroup/MoE-LLaVA.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/tmm.2026.3654458

Cites

20 of 20 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 20 references

Cited by

See all 69

Metadata

Authors
Lin, B., Tang, Z., Ye, Y., Huang, J., Zhang, J., Pang, Y., Jin, P., Ning, M., Luo, J., & Yuan, L.
Venue
IEEE Transactions on Multimedia
Published
2026
Volume
28
Pages
1-14
ISSN
  • 1520-9210 (ISSN-L)
  • 1941-0077
OpenAlex
W7124471359
Cited by
69
References
20
FWCI
105.95
Type
journal article
Language
en
Open access
green

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: MoE-LLaVA : Mixture of Experts for Large Vision-Language Models. https://app.cochat.ai/papers/10.1109/tmm.2026.3654458
The paper
Lin, B., Tang, Z., Ye, Y., Huang, J., Zhang, J., Pang, Y., Jin, P., Ning, M., Luo, J., & Yuan, L. (2026). MoE-LLaVA : Mixture of Experts for Large Vision-Language Models. IEEE Transactions on Multimedia. https://doi.org/10.1109/tmm.2026.3654458

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.