AndesVL Technical Report: An Efficient Mobile-side Multimodal Large Language Model
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915689469575168 |
|---|---|
| author | Jin, Zhiwei Song, Xiaohui Wang, Nan Liu, Yafei Li, Chao Li, Xin Wang, Ruichen Li, Zhihao Qi, Qi Cheng, Long Hao, Dongze Zheng, Quanlong Zhang, Yanhao Ji, Haobo Ma, Jian Zheng, Zhitong Lin, Zhenyi Deng, Haolin Zou, Xin Yin, Xiaojie Wang, Ruilin Cai, Liankai Liu, Haijing Qiu, Yuqing Chen, Ke Li, Zixian Xie, Chi Li, Huafei Li, Chenxing Wang, Chuangchuang Tang, Kai Zhu, Zhiguang Tang, Kai Gao, Wenmei Wang, Rui Wu, Jun Liu, Chao Xie, Qin Chen, Chen Lu, Haonan |
| author_facet | Jin, Zhiwei Song, Xiaohui Wang, Nan Liu, Yafei Li, Chao Li, Xin Wang, Ruichen Li, Zhihao Qi, Qi Cheng, Long Hao, Dongze Zheng, Quanlong Zhang, Yanhao Ji, Haobo Ma, Jian Zheng, Zhitong Lin, Zhenyi Deng, Haolin Zou, Xin Yin, Xiaojie Wang, Ruilin Cai, Liankai Liu, Haijing Qiu, Yuqing Chen, Ke Li, Zixian Xie, Chi Li, Huafei Li, Chenxing Wang, Chuangchuang Tang, Kai Zhu, Zhiguang Tang, Kai Gao, Wenmei Wang, Rui Wu, Jun Liu, Chao Xie, Qin Chen, Chen Lu, Haonan |
| contents | In recent years, while cloud-based MLLMs such as QwenVL, InternVL, GPT-4o, Gemini, and Claude Sonnet have demonstrated outstanding performance with enormous model sizes reaching hundreds of billions of parameters, they significantly surpass the limitations in memory, power consumption, and computing capacity of edge devices such as mobile phones. This paper introduces AndesVL, a suite of mobile-side MLLMs with 0.6B to 4B parameters based on Qwen3's LLM and various visual encoders. We comprehensively outline the model architectures, training pipeline, and training data of AndesVL, which achieves first-tier performance across a wide range of open-source benchmarks, including fields such as text-rich image understanding, reasoning and math, multi-image comprehension, general VQA, hallucination mitigation, multilingual understanding, and GUI-related tasks when compared with state-of-the-art models of a similar scale. Furthermore, we introduce a 1+N LoRA architecture alongside a Quantization-Aware LoRA Fine-Tuning (QALFT) framework to facilitate efficient task adaptation and model compression during mobile-side deployment of AndesVL. Moreover, utilizing our cache eviction algorithm -- OKV -- along with customized speculative decoding and compression strategies, we achieve a 6.7x peak decoding speedup ratio, up to 30.9% memory reduction, and 1.8 bits-per-weight when deploying AndesVL-4B on MediaTek Dimensity 9500 chips. We release all models on https://huggingface.co/OPPOer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_11496 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AndesVL Technical Report: An Efficient Mobile-side Multimodal Large Language Model Jin, Zhiwei Song, Xiaohui Wang, Nan Liu, Yafei Li, Chao Li, Xin Wang, Ruichen Li, Zhihao Qi, Qi Cheng, Long Hao, Dongze Zheng, Quanlong Zhang, Yanhao Ji, Haobo Ma, Jian Zheng, Zhitong Lin, Zhenyi Deng, Haolin Zou, Xin Yin, Xiaojie Wang, Ruilin Cai, Liankai Liu, Haijing Qiu, Yuqing Chen, Ke Li, Zixian Xie, Chi Li, Huafei Li, Chenxing Wang, Chuangchuang Tang, Kai Zhu, Zhiguang Tang, Kai Gao, Wenmei Wang, Rui Wu, Jun Liu, Chao Xie, Qin Chen, Chen Lu, Haonan Computer Vision and Pattern Recognition Artificial Intelligence In recent years, while cloud-based MLLMs such as QwenVL, InternVL, GPT-4o, Gemini, and Claude Sonnet have demonstrated outstanding performance with enormous model sizes reaching hundreds of billions of parameters, they significantly surpass the limitations in memory, power consumption, and computing capacity of edge devices such as mobile phones. This paper introduces AndesVL, a suite of mobile-side MLLMs with 0.6B to 4B parameters based on Qwen3's LLM and various visual encoders. We comprehensively outline the model architectures, training pipeline, and training data of AndesVL, which achieves first-tier performance across a wide range of open-source benchmarks, including fields such as text-rich image understanding, reasoning and math, multi-image comprehension, general VQA, hallucination mitigation, multilingual understanding, and GUI-related tasks when compared with state-of-the-art models of a similar scale. Furthermore, we introduce a 1+N LoRA architecture alongside a Quantization-Aware LoRA Fine-Tuning (QALFT) framework to facilitate efficient task adaptation and model compression during mobile-side deployment of AndesVL. Moreover, utilizing our cache eviction algorithm -- OKV -- along with customized speculative decoding and compression strategies, we achieve a 6.7x peak decoding speedup ratio, up to 30.9% memory reduction, and 1.8 bits-per-weight when deploying AndesVL-4B on MediaTek Dimensity 9500 chips. We release all models on https://huggingface.co/OPPOer. |
| title | AndesVL Technical Report: An Efficient Mobile-side Multimodal Large Language Model |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2510.11496 |