Better Supervised Fine-tuning for VQA: Integer-Only Loss

• URL: http://arxiv.org/abs/2508.11170v1
• Date: Fri, 15 Aug 2025 02:40:43 GMT
• Title: Better Supervised Fine-tuning for VQA: Integer-Only Loss
• Authors: Baihong Qian, Haotian Fan, Wenjie Liao, Yunqiu Wang, Tao Li, Junhui Cui,
• Abstract summary: IOVQA(Integer-only VQA) is a novel fine-tuning approach tailored for vision language models.<n>During dataset curation, we constrain the model's output to integers within the range of [10,50], ensuring numerical stability.<n>We also introduce a target-mask strategy: when computing the loss, only the first two-digit-integer of the label is unmasked.
• Score: 3.5053983607635533
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: With the rapid advancement of vision language models(VLM), their ability to assess visual content based on specific criteria and dimensions has become increasingly critical for applications such as video-theme consistency assessment and visual quality scoring. However, existing methods often suffer from imprecise results and inefficient loss calculation, which limit the focus of the model on key evaluation indicators. To address this, we propose IOVQA(Integer-only VQA), a novel fine-tuning approach tailored for VLMs to enhance their performance in video quality assessment tasks. The key innovation of IOVQA lies in its label construction and its targeted loss calculation mechanism. Specifically, during dataset curation, we constrain the model's output to integers within the range of [10,50], ensuring numerical stability, and convert decimal Overall_MOS to integer before using them as labels. We also introduce a target-mask strategy: when computing the loss, only the first two-digit-integer of the label is unmasked, forcing the model to learn the critical components of the numerical evaluation. After fine-tuning the Qwen2.5-VL model using the constructed dataset, experimental results demonstrate that the proposed method significantly improves the model's accuracy and consistency in the VQA task, ranking 3rd in VQualA 2025 GenAI-Bench AIGC Video Quality Assessment Challenge -- Track I. Our work highlights the effectiveness of merely leaving integer labels during fine-tuning, providing an effective idea for optimizing VLMs in quantitative evaluation scenarios.

Related papers

• Q-Save: Towards Scoring and Attribution for Generated Video Evaluation [65.83319736145869]
  We present Q-Save, a new benchmark dataset and model for holistic evaluation of AI-generated video (AIGV) quality.<n>The dataset contains near 10000 videos, each annotated with a scalar mean opinion score (MOS) and fine-grained attribution labels.<n>We propose a unified evaluation model that jointly performs quality scoring and attribution-based explanation.
  arXiv  Detail & Related papers  (2025-11-24T07:00:21Z)
• CAMP-VQA: Caption-Embedded Multimodal Perception for No-Reference Quality Assessment of Compressed Video [9.172799792564009]
  We propose CAMP-VQA, a novel NR-VQA framework that exploits the semantic understanding capabilities of large models.<n>Our approach introduces a quality-aware video metadata mechanism that integrates key fragments extracted from inter-frame variations.<n>Our model consistently outperforms existing NR-VQA methods, achieving improved accuracy without the need for costly manual fine-grained annotations.
  arXiv  Detail & Related papers  (2025-11-10T16:37:47Z)
• VQAThinker: Exploring Generalizable and Explainable Video Quality Assessment via Reinforcement Learning [50.34205095371895]
  Video quality assessment aims to objectively quantify perceptual quality degradation.<n>Existing VQA models suffer from two critical limitations.<n>We propose textbfVQAThinker, a reasoning-based VQA framework.
  arXiv  Detail & Related papers  (2025-08-08T06:16:23Z)
• Breaking Annotation Barriers: Generalized Video Quality Assessment via Ranking-based Self-Supervision [49.46606936180063]
  Video quality assessment (VQA) is essential for quantifying quality in various video processing systems.<n>We introduce a self-supervised learning framework for VQA to learn quality assessment capabilities from large-scale, unlabeled web videos.<n>By training on a dataset $10times$ larger than the existing VQA benchmarks, our model achieves zero-shot performance.
  arXiv  Detail & Related papers  (2025-05-06T15:29:32Z)
• Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success [100.226572152954]
  We present an optimized fine-tuning recipe for vision-language-action models (VLAs)<n>Our recipe boosts OpenVLA's average success rate across four task suites from 76.5% to 97.1% while increasing action generation throughput by 26$times$.<n>In real-world evaluations, our fine-tuning recipe enables OpenVLA to successfully execute dexterous, high-frequency control tasks on a bimanual ALOHA robot.
  arXiv  Detail & Related papers  (2025-02-27T00:30:29Z)
• VQA$^2$: Visual Question Answering for Video Quality Assessment [76.81110038738699]
  Video Quality Assessment (VQA) is a classic field in low-level visual perception.<n>Recent studies in the image domain have demonstrated that Visual Question Answering (VQA) can enhance markedly low-level visual quality evaluation.<n>We introduce the VQA2 Instruction dataset - the first visual question answering instruction dataset that focuses on video quality assessment.<n>The VQA2 series models interleave visual and motion tokens to enhance the perception of spatial-temporal quality details in videos.
  arXiv  Detail & Related papers  (2024-11-06T09:39:52Z)
• VHELM: A Holistic Evaluation of Vision Language Models [75.88987277686914]
  We present the Holistic Evaluation of Vision Language Models (VHELM) VHELM aggregates various datasets to cover one or more of the 9 aspects: visual perception, knowledge, reasoning, bias, fairness, multilinguality, robustness, toxicity, and safety. Our framework is designed to be lightweight and automatic so that evaluation runs are cheap and fast.
  arXiv  Detail & Related papers  (2024-10-09T17:46:34Z)
• Charting the Future: Using Chart Question-Answering for Scalable Evaluation of LLM-Driven Data Visualizations [7.32619928577074]
  We propose a novel framework that leverages Visual Question Answering (VQA) models to automate the evaluation of LLM-generated data visualizations. Our results indicate that LLM-generated charts do not match the accuracy of the original non-LLM-generated charts based on VQA performance measures.
  arXiv  Detail & Related papers  (2024-09-27T14:02:48Z)
• Analysis of Video Quality Datasets via Design of Minimalistic Video Quality Models [71.06007696593704]
  Blind quality assessment (BVQA) plays an indispensable role in monitoring and improving the end-users' viewing experience in real-world video-enabled media applications. As an experimental field, the improvements of BVQA models have been measured primarily on a few human-rated VQA datasets. We conduct a first-of-its-kind computational analysis of VQA datasets via minimalistic BVQA models.
  arXiv  Detail & Related papers  (2023-07-26T06:38:33Z)
• Declaration-based Prompt Tuning for Visual Question Answering [16.688288454811016]
  We propose an innovative visual-language (VL) fine-tuning paradigm (named Declaration-based Prompt Tuning, abbreviated as DPT) DPT jointly optimize the objectives of pre-training and fine-tuning of VQA model, boosting the effective adaptation of pre-trained VL models to the downstream task. Experimental results on GQA dataset show that DPT outperforms the fine-tuned counterpart by a large margin regarding accuracy in both fully-supervised (2.68%) and zero-shot/few-shot (over 31%) settings.
  arXiv  Detail & Related papers  (2022-05-05T05:56:55Z)
• UGC-VQA: Benchmarking Blind Video Quality Assessment for User Generated Content [59.13821614689478]
  Blind quality prediction of in-the-wild videos is quite challenging, since the quality degradations of content are unpredictable, complicated, and often commingled. Here we contribute to advancing the problem by conducting a comprehensive evaluation of leading VQA models. By employing a feature selection strategy on top of leading VQA model features, we are able to extract 60 of the 763 statistical features used by the leading models. Our experimental results show that VIDEVAL achieves state-of-theart performance at considerably lower computational cost than other leading models.
  arXiv  Detail & Related papers  (2020-05-29T00:39:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.