Salvato in:
Dettagli Bibliografici
Autori principali: Li, Siqi, Shen, Yufan, Chen, Xiangnan, Chen, Jiayi, Ju, Hengwei, Duan, Haodong, Mao, Song, Zhou, Hongbin, Zhang, Bo, Fu, Bin, Cai, Pinlong, Wen, Licheng, Shi, Botian, Liu, Yong, Cai, Xinyu, Qiao, Yu
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2505.00063
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918029674151936
author Li, Siqi
Shen, Yufan
Chen, Xiangnan
Chen, Jiayi
Ju, Hengwei
Duan, Haodong
Mao, Song
Zhou, Hongbin
Zhang, Bo
Fu, Bin
Cai, Pinlong
Wen, Licheng
Shi, Botian
Liu, Yong
Cai, Xinyu
Qiao, Yu
author_facet Li, Siqi
Shen, Yufan
Chen, Xiangnan
Chen, Jiayi
Ju, Hengwei
Duan, Haodong
Mao, Song
Zhou, Hongbin
Zhang, Bo
Fu, Bin
Cai, Pinlong
Wen, Licheng
Shi, Botian
Liu, Yong
Cai, Xinyu
Qiao, Yu
contents The rapid advancement of multimodal large language models (MLLMs) has profoundly impacted the document domain, creating a wide array of application scenarios. This progress highlights the need for a comprehensive benchmark to evaluate these models' capabilities across various document-specific tasks. However, existing benchmarks often fail to locate specific model weaknesses or guide systematic improvements. To bridge this gap, we introduce a General Document Intelligence Benchmark (GDI-Bench), featuring 2.3k images across 9 key scenarios and 19 document-specific tasks. By decoupling visual complexity and reasoning complexity, the GDI-Bench structures graded tasks that allow performance assessment by difficulty, aiding in model weakness identification and optimization guidance. We evaluate various open-source and closed-source models on GDI-Bench, conducting decoupled analyses in the visual and reasoning domains, revealing their strengths and weaknesses. To address the diverse tasks and domains in the GDI-Bench, we propose a GDI-Model that mitigates catastrophic forgetting during the supervised fine-tuning (SFT) process through an intelligence-preserving training strategy, thereby reinforcing the inherent weaknesses of the base model. Our model achieves state-of-the-art performance on previous benchmarks and the GDI-Bench. Both our benchmark and models are or will be open-sourced on https://huggingface.co/GDIBench.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00063
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GDI-Bench: A Benchmark for General Document Intelligence with Vision and Reasoning Decoupling
Li, Siqi
Shen, Yufan
Chen, Xiangnan
Chen, Jiayi
Ju, Hengwei
Duan, Haodong
Mao, Song
Zhou, Hongbin
Zhang, Bo
Fu, Bin
Cai, Pinlong
Wen, Licheng
Shi, Botian
Liu, Yong
Cai, Xinyu
Qiao, Yu
Computation and Language
Computer Vision and Pattern Recognition
The rapid advancement of multimodal large language models (MLLMs) has profoundly impacted the document domain, creating a wide array of application scenarios. This progress highlights the need for a comprehensive benchmark to evaluate these models' capabilities across various document-specific tasks. However, existing benchmarks often fail to locate specific model weaknesses or guide systematic improvements. To bridge this gap, we introduce a General Document Intelligence Benchmark (GDI-Bench), featuring 2.3k images across 9 key scenarios and 19 document-specific tasks. By decoupling visual complexity and reasoning complexity, the GDI-Bench structures graded tasks that allow performance assessment by difficulty, aiding in model weakness identification and optimization guidance. We evaluate various open-source and closed-source models on GDI-Bench, conducting decoupled analyses in the visual and reasoning domains, revealing their strengths and weaknesses. To address the diverse tasks and domains in the GDI-Bench, we propose a GDI-Model that mitigates catastrophic forgetting during the supervised fine-tuning (SFT) process through an intelligence-preserving training strategy, thereby reinforcing the inherent weaknesses of the base model. Our model achieves state-of-the-art performance on previous benchmarks and the GDI-Bench. Both our benchmark and models are or will be open-sourced on https://huggingface.co/GDIBench.
title GDI-Bench: A Benchmark for General Document Intelligence with Vision and Reasoning Decoupling
topic Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.00063