AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

EVEv2: Improved Baselines for Encoder-Free Vision-Language Models

Haiwen Diao, Xiaotong Li, Yufeng Cui, Yueze Wang, Haoge Deng, Ting Pan, Wenxuan Wang, Huchuan Lu, Xinlong Wang

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis · 33 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
fully decouple attention, LayerNorm and FFN per modality (vs decoupling only LayerNorm) (ablation)EVEv2.0 (Qwen2.5-7B) · Avg. over benchmarks, 8M dataAvg. accuracy: 48.8% -> 51.6%Improved
use Divide-and-Conquer (full decoupling) instead of MoE (FFN-only) as data grows (ablation)EVEv2.0 vs EVEv1.5 · Avg. over benchmarks, 8M to 24M dataAvg. accuracy gap: 0.8% -> 1.4%Improved
full modality decoupling vs MoE vs dense on Qwen2.5-1.5B (ablation)EVE variants · Six VL benchmarks, EVE-recap-4MAvg. score: 35.1% -> 37.3% -> 38.5%Improved
full decoupling vs MoE vs ReP (SFT loss) (ablation)EVEv1.2, EVEv1.5, EVEv2 · SFT stageSFT loss: 1.0306 -> 0.9513 -> 0.9327Improved
re-parameterization (update LLM weights) instead of frozen LLM with vision layers (ablation)EVEv1.2 · SQA-IMGaccuracyHurt
add modality-specific weights (EVEv2.0) vs EVEv1.5 MoE (ablation)EVEv1.5 · All multimodal benchmarks, Stage 2accuracyImproved
recaption data with DenseFusion++ (7B) vs LLaVA-1.5 (13B)/Emu2 (17B) captions or raw web captions (ablation)EVEv1.0 · Pretraining data comparisonbenchmark accuracyImproved
mix LAION, OpenImages and SAM data into pretraining (ablation)EVEv1.0 · Pretraining data mixturebenchmark accuracyImproved
use AnyRatio_HD (fixed 1600x1600 area) image processing (ablation)EVEv1.0 (Vicuna-7B) · Image input formatsbenchmark accuracyImproved
use AnyResolution input (ablation)EVEv1.0 (Vicuna-7B) · Image input formatsbenchmark accuracyMixed
add modality-specific weights (EVEv2.0 vs EVEv1.0 dense) (ablation)EVEv1.0 · TTFTtime to first token (s): 3.77 -> 4.25Hurt
unfreeze LLM when increasing captioning data (ablation)EVE variants · SQA-IMGaccuracyHurt

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