AI Paper Atlas

Research map / Point clouds / Point cloud classification

PointLLM-V2: Empowering Large Language Models to Better Understand Point Clouds

Runsen Xu, Shuai Yang, Xiaolong Wang, Tai Wang, Yilun Chen, Jiangmiao Pang, Dahua Lin

TPAMI 2025 · Point clouds › Point cloud classification · approach: Point-language and open-vocabulary 3D · 12 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use a 3-layer projector (hidden 1024, 2048) instead of 1 layer (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: 51.05 -> 52.82Improved
use a 3-layer projector instead of 1 layer (ablation)PointLLM-13B · Generative classification average accuracyaccuracy: 49.00 -> 53.39Improved
use a 3-layer projector instead of a 4-layer one (ablation)PointLLM-13B · Generative classification average accuracyaccuracy: 51.40 -> 53.39Improved
use a 3-layer projector instead of 4 layers (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: 52.15 -> 52.82Improved
add max pooling of point tokens before the projector (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: 52.82 -> 48.72Hurt
add max pooling of point tokens before the projector (ablation)PointLLM-13B · Generative classification average accuracyaccuracy: 53.39 -> 51.10Hurt
add max pooling of point tokens (ablation)PointLLM-7B · Training costA100 GPU-hours: 126 -> 34Improved
add max pooling of point tokens (ablation)PointLLM-13B · Training costA100 GPU-hours: 213 -> 56Improved
add multi-round conversations to fine-tuning data (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: 40.14 -> 45.79Improved
add detailed descriptions to fine-tuning data (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: 45.79 -> 52.82Improved
increase alignment-stage data volume (ablation)PointLLM-7B · Generative classification average accuracyaccuracy: plateaus at around 600K samplesImproved
prompt for captions of at most 20 wordsPointLLM-13B · 3D object captioning (Objaverse)BLEU-1: 3.83 -> 17.09Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy