Explorando o Treinamento Autoevolutivo para Raciocínio Multimodal
Diving into Self-Evolving Training for Multimodal Reasoning
December 23, 2024
Autores: Wei Liu, Junlong Li, Xiwen Zhang, Fan Zhou, Yu Cheng, Junxian He
cs.AI
Resumo
A capacidade de raciocínio é essencial para os Modelos Multimodais de Grande Escala (LMMs). Na ausência de dados anotados de corrente de pensamento multimodal, o treinamento autoevolutivo, no qual o modelo aprende a partir de suas próprias saídas, surgiu como uma abordagem eficaz e escalável para aprimorar as habilidades de raciocínio. Apesar de seu uso crescente, a compreensão abrangente do treinamento autoevolutivo, especialmente no contexto do raciocínio multimodal, ainda é limitada. Neste artigo, aprofundamos nas complexidades do treinamento autoevolutivo para o raciocínio multimodal, destacando três fatores-chave: Método de Treinamento, Modelo de Recompensa e Variação de Estímulo. Examinamos sistematicamente cada fator e exploramos como várias configurações afetam a eficácia do treinamento. Nossa análise resulta em um conjunto das melhores práticas para cada fator, com o objetivo de otimizar o raciocínio multimodal. Além disso, exploramos a Dinâmica de Autoevolução durante o treinamento e o impacto dos mecanismos automáticos de equilíbrio no aumento de desempenho. Após todas as investigações, apresentamos uma receita final para o treinamento autoevolutivo no raciocínio multimodal, encapsulando essas escolhas de design em um framework que chamamos de MSTaR (Treinamento Autoevolutivo Multimodal para Raciocínio), que é universalmente eficaz para modelos de diferentes tamanhos em vários benchmarks, por exemplo, superando significativamente o modelo pré-evoluído em 5 benchmarks de raciocínio multimodal sem o uso de anotações humanas adicionais, conforme demonstrado no MiniCPM-V-2.5 (8B), Phi-3.5-Vision (4B) e InternVL2 (2B). Acreditamos que este estudo preenche uma lacuna significativa na compreensão do treinamento autoevolutivo para o raciocínio multimodal e oferece um framework robusto para pesquisas futuras. Nossos modelos de política e recompensa, bem como os dados coletados, são disponibilizados para facilitar investigações adicionais no raciocínio multimodal.
English
Reasoning ability is essential for Large Multimodal Models (LMMs). In the
absence of multimodal chain-of-thought annotated data, self-evolving training,
where the model learns from its own outputs, has emerged as an effective and
scalable approach for enhancing reasoning abilities. Despite its growing usage,
a comprehensive understanding of self-evolving training, particularly in the
context of multimodal reasoning, remains limited. In this paper, we delve into
the intricacies of self-evolving training for multimodal reasoning, pinpointing
three key factors: Training Method, Reward Model, and Prompt Variation. We
systematically examine each factor and explore how various configurations
affect the training's effectiveness. Our analysis leads to a set of best
practices for each factor, aimed at optimizing multimodal reasoning.
Furthermore, we explore the Self-Evolution Dynamics during training and the
impact of automatic balancing mechanisms in boosting performance. After all the
investigations, we present a final recipe for self-evolving training in
multimodal reasoning, encapsulating these design choices into a framework we
call MSTaR (Multimodal Self-evolving Training for Reasoning), which is
universally effective for models with different sizes on various benchmarks,
e.g., surpassing the pre-evolved model significantly on 5 multimodal reasoning
benchmarks without using additional human annotations, as demonstrated on
MiniCPM-V-2.5 (8B), Phi-3.5-Vision (4B) and InternVL2 (2B). We believe this
study fills a significant gap in the understanding of self-evolving training
for multimodal reasoning and offers a robust framework for future research. Our
policy and reward models, as well as the collected data, is released to
facilitate further investigation in multimodal reasoning.Summary
AI-Generated Summary
DeepSeek-R1: Incentivizando a Capacidade de Raciocínio em LLMs via
Aprendizado por ReforçoDeepSeek-R1: Incentivizing Reasoning Capability in LLMs via
Reinforcement Learning
DeepSeek-R1: Incentivizando a Capacidade de Raciocínio em LLMs via
Aprendizado por Reforço
DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via
Reinforcement Learning
DeepSeek-AI, Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, Xiaokang Zhang, Xingkai Yu, Yu Wu, Z. F. Wu, Zhibin Gou, Zhihong Shao, Zhuoshu Li, Ziyi Gao, Aixin Liu, Bing Xue, Bingxuan Wang, Bochao Wu, Bei Feng, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, Damai Dai, Deli Chen, Dongjie Ji, Erhang Li, Fangyun Lin, Fucong Dai, Fuli Luo, Guangbo Hao, Guanting Chen, Guowei Li, H. Zhang, Han Bao, Hanwei Xu, Haocheng Wang, Honghui Ding, Huajian Xin, Huazuo Gao, Hui Qu, Hui Li, Jianzhong Guo, Jiashi Li, Jiawei Wang, Jingchang Chen, Jingyang Yuan, Junjie Qiu, Junlong Li, J. L. Cai, Jiaqi Ni, Jian Liang, Jin Chen, Kai Dong, Kai Hu, Kaige Gao, Kang Guan, Kexin Huang, Kuai Yu, Lean Wang, Lecong Zhang, Liang Zhao, Litong Wang, Liyue Zhang, Lei Xu, Leyi Xia, Mingchuan Zhang, Minghua Zhang, Minghui Tang, Meng Li, Miaojun Wang, Mingming Li, Ning Tian, Panpan Huang, Peng Zhang, Qiancheng Wang, Qinyu Chen, Qiushi Du, Ruiqi Ge, Ruisong Zhang, Ruizhe Pan, Runji Wang, R. J. Chen, R. L. Jin, Ruyi Chen, Shanghao Lu, Shangyan Zhou, Shanhuang Chen, Shengfeng Ye, Shiyu Wang, Shuiping Yu, Shunfeng Zhou, Shuting Pan, S. S. Li, Shuang Zhou, Shaoqing Wu, Shengfeng Ye, Tao Yun, Tian Pei, Tianyu Sun, T. Wang, Wangding Zeng, Wanjia Zhao, Wen Liu, Wenfeng Liang, Wenjun Gao, Wenqin Yu, Wentao Zhang, W. L. Xiao, Wei An, Xiaodong Liu, Xiaohan Wang, Xiaokang Chen, Xiaotao Nie, Xin Cheng, Xin Liu, Xin Xie, Xingchao Liu, Xinyu Yang, Xinyuan Li, Xuecheng Su, Xuheng Lin, X. Q. Li, Xiangyue Jin, Xiaojin Shen, Xiaosha Chen, Xiaowen Sun, Xiaoxiang Wang, Xinnan Song, Xinyi Zhou, Xianzu Wang, Xinxia Shan, Y. K. Li, Y. Q. Wang, Y. X. Wei, Yang Zhang, Yanhong Xu, Yao Li, Yao Zhao, Yaofeng Sun, Yaohui Wang, Yi Yu, Yichao Zhang, Yifan Shi, Yiliang Xiong, Ying He, Yishi Piao, Yisong Wang, Yixuan Tan, Yiyang Ma, Yiyuan Liu, Yongqiang Guo, Yuan Ou, Yuduan Wang, Yue Gong, Yuheng Zou, Yujia He, Yunfan Xiong, Yuxiang Luo, Yuxiang You, Yuxuan Liu, Yuyang Zhou, Y. X. Zhu, Yanhong Xu, Yanping Huang, Yaohui Li, Yi Zheng, Yuchen Zhu, Yunxian Ma, Ying Tang, Yukun Zha, Yuting Yan, Z. Z. Ren, Zehui Ren, Zhangli Sha, Zhe Fu, Zhean Xu, Zhenda Xie, Zhengyan Zhang, Zhewen Hao, Zhicheng Ma, Zhigang Yan, Zhiyu Wu, Zihui Gu, Zijia Zhu, Zijun Liu, Zilin Li, Ziwei Xie, Ziyang Song, Zizheng Pan, Zhen Huang, Zhipeng Xu, Zhongyu Zhang, Zhen Zhang•Jan 22, 2025•3735
Relatório Técnico do Qwen2.5Qwen2.5 Technical Report
Relatório Técnico do Qwen2.5
Qwen2.5 Technical Report
Qwen, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, Keming Lu, Keqin Bao, Kexin Yang, Le Yu, Mei Li, Mingfeng Xue, Pei Zhang, Qin Zhu, Rui Men, Runji Lin, Tianhao Li, Tingyu Xia, Xingzhang Ren, Xuancheng Ren, Yang Fan, Yang Su, Yichang Zhang, Yu Wan, Yuqiong Liu, Zeyu Cui, Zhenru Zhang, Zihan Qiu•Dec 19, 2024•36311
MiniMax-01: Dimensionamento de Modelos de Fundação com Atenção RelâmpagoMiniMax-01: Scaling Foundation Models with Lightning Attention
MiniMax-01: Dimensionamento de Modelos de Fundação com Atenção Relâmpago
MiniMax-01: Scaling Foundation Models with Lightning Attention
MiniMax, Aonian Li, Bangwei Gong, Bo Yang, Boji Shan, Chang Liu, Cheng Zhu, Chunhao Zhang, Congchao Guo, Da Chen, Dong Li, Enwei Jiao, Gengxin Li, Guojun Zhang, Haohai Sun, Houze Dong, Jiadai Zhu, Jiaqi Zhuang, Jiayuan Song, Jin Zhu, Jingtao Han, Jingyang Li, Junbin Xie, Junhao Xu, Junjie Yan, Kaishun Zhang, Kecheng Xiao, Kexi Kang, Le Han, Leyang Wang, Lianfei Yu, Liheng Feng, Lin Zheng, Linbo Chai, Long Xing, Meizhi Ju, Mingyuan Chi, Mozhi Zhang, Peikai Huang, Pengcheng Niu, Pengfei Li, Pengyu Zhao, Qi Yang, Qidi Xu, Qiexiang Wang, Qin Wang, Qiuhui Li, Ruitao Leng, Shengmin Shi, Shuqi Yu, Sichen Li, Songquan Zhu, Tao Huang, Tianrun Liang, Weigao Sun, Weixuan Sun, Weiyu Cheng, Wenkai Li, Xiangjun Song, Xiao Su, Xiaodong Han, Xinjie Zhang, Xinzhu Hou, Xu Min, Xun Zou, Xuyang Shen, Yan Gong, Yingjie Zhu, Yipeng Zhou, Yiran Zhong, Yongyi Hu, Yuanxiang Fan, Yue Yu, Yufeng Yang, Yuhao Li, Yunan Huang, Yunji Li, Yunpeng Huang, Yunzhi Xu, Yuxin Mao, Zehan Li, Zekang Li, Zewei Tao, Zewen Ying, Zhaoyang Cong, Zhen Qin, Zhenhua Fan, Zhihang Yu, Zhuo Jiang, Zijia Wu•Jan 14, 2025•2836