paper-with-me

Papers

TR-PTS: Task-Relevant Parameter and Token Selection for Efficient Tuning

2025-07-30 · Siqi Luo, Haoran Yang, Yi Xin, Mingyang Yi, Guangyang Wu, Guangtao Zhai, Xiaohong Liu arxiv

Large pre-trained models achieve remarkable performance in vision tasks but are impractical for fine-tuning due to high computational and storage costs. Parameter-Efficient Fine-Tuning (PEFT) methods mitigate this issue by updating only a subset of parameters; however, most existing approaches are task-agnostic, failing to fully exploit task-specific adaptations, which leads to suboptimal efficiency and performance. To address this limitation, we propose Task-Relevant Parameter and Token Selection (TR-PTS), a task-driven framework that enhances both computational efficiency and accuracy. Specifically, we introduce Task-Relevant Parameter Selection, which utilizes the Fisher Information Matrix (FIM) to identify and fine-tune only the most informative parameters in a layer-wise manner, while keeping the remaining parameters frozen. Simultaneously, Task-Relevant Token Selection dynamically preserves the most informative tokens and merges redundant ones, reducing computational overhead. By jointly optimizing parameters and tokens, TR-PTS enables the model to concentrate on task-discriminative information. We evaluate TR-PTS on benchmark, including FGVC and VTAB-1k, where it achieves state-of-the-art performance, surpassing full fine-tuning by 3.40% and 10.35%, respectively. The code are available at https://github.com/synbol/TR-PTS.

📄 PDF Abstract BibTeX arXiv:2507.22872

Code (0)

등록된 구현이 없습니다.

Tasks

parameter-efficient fine-tuningComputational Efficiency

Similar Papers 제목 키워드 기반

Spotlighter: Revisiting Prompt Tuning from a Representative Mining View

2025-08-31 · Yutong Gao, Maoyuan Shao, Xinyang Huang, Chuang Zhu 외 arxiv

CLIP's success has demonstrated that prompt tuning can achieve robust cross-modal semantic alignment for tasks ranging from open-domain recognition to fine-grained classification. However, redundant or weakly relevant fe…

Classifier Language Models: Unifying Sparse Finetuning and Adaptive Tokenization for Specialized Classification Tasks

2025-08-12 · Adit Krishnan, Chu Wang, Chris Kong arxiv

Semantic text classification requires the understanding of the contextual significance of specific tokens rather than surface-level patterns or keywords (as in rule-based or statistical text classification), making large…

Semantic Role LabelingText ClassificationIntent Detection

Efficient Multi-View 3D Object Detection by Dynamic Token Selection and Fine-Tuning

2026-04-15 · Danish Nazir, Antoine Hanna-Asaad, Lucas Görnhardt, Jan Piewek 외 arxiv

Existing multi-view three-dimensional (3D) object detection approaches widely adopt large-scale pre-trained vision transformer (ViT)-based foundation models as backbones, being computationally complex. To address this pr…

parameter-efficient fine-tuningComputational Efficiency3D Object Detection

PointLoRA: Low-Rank Adaptation with Token Selection for Point Cloud Learning

2025-04-22 · CVPR 2025 1 · Song Wang, Xiaolu Liu, Lingdong Kong, Jianyun Xu 외

Self-supervised representation learning for point cloud has demonstrated effectiveness in improving pre-trained model performance across diverse tasks. However, as pre-trained models grow in complexity, fully fine-tuning…

parameter-efficient fine-tuningRepresentation Learning

Employing an Adjusted Stability Measure for Multi-Criteria Model Fitting on Data Sets with Similar Features

2021-06-15 · Andrea Bommert, Jörg Rahnenführer, Michel Lang

Fitting models with high predictive accuracy that include all relevant but no irrelevant or redundant features is a challenging task on data sets with similar (e.g. highly correlated) features. We propose the approach of…

feature selection