SOURCE-LINKED INTELLIGENCE

Position Anchor Tuning: Towards Efficient Adaptation of Pre-Trained Point Cloud Transformers

arXiv · AI, language, vision and robotics · article · Sep 16, 2026 · UTC

Parameter-efficient fine-tuning (PEFT) has recently emerged as a pivotal research direction for adapting pre-trained point cloud transformers to diverse downstream tasks. Although existing methods achieve excellent fine-tuning performance with high parameter efficiency, they ignore inference efficiency. To tackle this problem, a novel PEFT method termed position anchor tuning (PAT) is proposed in this paper. As multi-head attention (MHA) and feed-forward network (FFN) are computation-heavy blocks in pre-trained transformers, PAT decreases their computational cost through token aggregation-expa

Read original source ↗ Open in workspace

recordType

paper

region

Global