AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

MUSE-VL: Modeling Unified VLM through Semantic Discrete Encoding

Rongchang Xie, Chen Du, Ping Song, Chang Liu

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning · 30 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace VQGAN tokenizer with semantic SDE tokenizerYi-1.5-9B VLM · MMBench/SEED/MMStar averageaverage accuracy: 34.6 -> 60.8Improved
use SDE tokenizer instead of SEED tokenizerYi-1.5-9B VLM · MMBench/SEED/MMStar averageaverage accuracy: 53.3 -> 60.8Improved
use SDE tokenizer instead of LaVIT tokenizerYi-1.5-9B VLM · MMBench/SEED/MMStar averageaverage accuracy: 54.4 -> 60.8Improved
add semantic branch (SigLIP feature reconstruction) to image-only tokenizer (ablation)SDE tokenizer · MMB/SEED/MMStar averageaverage accuracy: 43.1 -> 63.6Improved
add semantic branch to image-only tokenizer (ablation)SDE tokenizer · ImageNet val reconstructionrFID: 2.63 -> 2.26Improved
add image reconstruction branch to semantic-only tokenizer (ablation)SDE tokenizer · MMB/SEED/MMStar averageaverage accuracy: 62.7 -> 63.6Improved
increase input resolution from 256 to 384 (ablation)MUSE-VL Yi-1.5-9B · MMB/SEED/MMStar averageaverage accuracy: 60.2 -> 63.3Improved
scale LLM from Yi-1.5-9B to Yi-1.5-34B (ablation)MUSE-VL at 256 res · MMB/SEED/MMStar averageaverage accuracy: 60.2 -> 63.2Improved
scale LLM from Qwen-2.5-7B to Qwen-2.5-32B (ablation)MUSE-VL at 256 res · MMB/SEED/MMStar averageaverage accuracy: 60.3 -> 63.7Improved
scale LLM from Qwen-2.5-7B to Qwen-2.5-32B (ablation)MUSE-VL · SEED-Bench-img (ablation subset)accuracy: 65.8 -> 65.7No clear effect
prompt rewriting for text-to-image generationMUSE-VL 7B · GenEvaloverall score: 0.53 -> 0.57Improved
scale MUSE-VL from 7B to 32BMUSE-VL · Understanding average (7 benchmarks)average accuracy: 63.6 -> 70.1Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy