AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

FlashVLM: Text-Guided Visual Token Selection for Large Multimodal Models

Kaitong Cai, jusheng zhang, Sizhuo Ma, Bingqian Lu, Jian Wang, Keze Wang

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Training-free token reduction for MLLMs · 15 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
add text-guided extrinsic similarity (ablation)FlashVLM · VQAv2 (64 tokens)accuracy: 72.7 -> 73.6Improved
add intrinsic visual saliency (ablation)FlashVLM · VQAv2 (64 tokens)accuracy: 72.5 -> 73.6Improved
add diversity-preserving background reserve (ablation)FlashVLM · VQAv2 (64 tokens)accuracy: 72.4 -> 73.6Improved
use log-domain fusion instead of linear fusion (ablation)FlashVLM · POPE (64 tokens)accuracy: 80.2 -> 81.7Improved
add text sharpening (ablation)FlashVLM · VQAv2 (64 tokens)accuracy: 73.0 -> 73.6Improved
add query gating (ablation)FlashVLM · VQAv2 (64 tokens)accuracy: 72.8 -> 73.6Improved
add text guidance at 128 tokens (ablation)FlashVLM · VQAv2 (128 tokens)accuracy: 75.9 -> 76.4Improved
add diversity reserve at 128 tokens (ablation)FlashVLM · GQA (128 tokens)accuracy: 58.1 -> 58.9Improved
prune 77.8% of visual tokens (keep 128) with FlashVLMLLaVA-1.5-7B · 10 image benchmarksrelative accuracy: 100.0% -> 100.6%Improved
increase compression from 128 to 32 tokensFlashVLM · 10 image benchmarksrelative accuracy: 100.6% -> 92.8%Hurt
replace VisPruner with FlashVLM at 32 tokensVisPruner · 10 image benchmarksrelative accuracy: 91.5% -> 92.8%Improved
replace VisPruner with FlashVLM at 455 tokensVisPruner · Video QA avgaccuracy: 48.4 -> 48.7Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 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