Q-Adapt: Adapting LMM for Visual Quality Assessment with Progressive Instruction Tuning
Yiting Lu, Xin Li, Haoning Wu, Bingchen Li, Weisi Lin, Zhibo Chen, [1mm] University of Science, Technology of China, Nanyang Technological University {luyt31415,lbc31415926}@mail.ustc.edu.cn, [email protected],[email protected], {xin.li, chenzhibo}@ustc.edu.cn · 2025
Read Q-Adapt: Adapting LMM for Visual Quality Assessment with Progressive Instruction Tuning by Yiting Lu, Xin Li, Haoning Wu, Bingchen Li, Weisi Lin, Zhibo Chen, [1mm] University of Science, Technology of China, Nanyang Technological University {luyt31415,lbc31415926}@mail.ustc.edu.cn, [email protected],[email protected], {xin.li, chenzhibo}@ustc.edu.cn (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-01655.txt.
Description
Yiting Lu and co-authors present Q-Adapt in 2025, a way to teach large multimodal models to judge image quality. A good judge must give an overall verdict and answer questions about single traits like sharpness or noise. Training both skills at once creates conflicts.
Q-Adapt splits training into two stages. First, a light adapter gives the model shared perception knowledge for both tasks. Then visual prompt tuning adjusts image features to fit each instruction type. The result is a lightweight evaluator that explains its judgments. It matched or beat rivals on standard quality benchmarks and databases. The approach suits tools that must explain, not just score, picture quality.
Catalog record
| CREATOR | Yiting Lu, Xin Li, Haoning Wu, Bingchen Li, Weisi Lin, Zhibo Chen, [1mm] University of Science, Technology of China, Nanyang Technological University {luyt31415,lbc31415926}@mail.ustc.edu.cn, [email protected],[email protected], {xin.li, chenzhibo}@ustc.edu.cn |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-01655