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

ICCV 2025poster0 citations

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

Abstract

Existing encoder-free vision-language models (VLMs) are rapidly narrowing the performance gap with their encoder-based counterparts, highlighting the promising potential for unified multimodal systems with structural simplicity and efficient deployment. We systematically clarify the performance gap between VLMs using pre-trained vision encoders, discrete tokenizers, and minimalist visual layers from scratch, deeply excavating the under-examined characteristics of encoder-free VLMs. We develop efficient strategies for encoder-free VLMs that rival mainstream encoder-based ones. After an in-depth investigation, we launch EVEv2.0, a new and improved family of encoder-free VLMs. We show that: (i) Properly decomposing and hierarchically associating vision and language within a unified model reduces interference between modalities. (ii) A well-designed training strategy enables effective optimization for encoder-free VLMs. Through extensive evaluation, our EVEv2.0 represents a thorough study for developing a decoder-only architecture across modalities, demonstrating superior data efficiency and strong vision-reasoning capability. Code is publicly available at: https://github.com/baaivision/EVE.

BibTeX

@InProceedings{Diao_2025_ICCV,
    author    = {Diao, Haiwen and Li, Xiaotong and Cui, Yufeng and Wang, Yueze and Deng, Haoge and Pan, Ting and Wang, Wenxuan and Lu, Huchuan and Wang, Xinlong},
    title     = {EVEv2: Improved Baselines for Encoder-Free Vision-Language Models},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2025},
    pages     = {21014-21025}
}