WebJul 3, 2024 · PaddlePaddle / PaddleVideo Public. #547 opened on Oct 24, 2024 by zpc-666 Loading…. Add STGCN++ model. #533 opened on Sep 21, 2024 by txyugood Loading…. #506 opened on Aug 12, 2024 by Wuxiao85 Loading…. #501 opened on Aug 10, 2024 by Aganlengzi Loading…. #487 opened on Jul 20, 2024 by Gary2024X Loading…. 2.1 CTRGCN_light的简述 最初我们是一味地使用轻量的模块如ghost模块来代替原始卷积从而减少总的参数量和浮点数运算量,但无疑增加了模型的复杂性,模型结构变得更加琐碎,这使得模型权重文件(即CTRGCN_joint.pdiparams)得以大幅度减小,但带来的代价是模型结构文件大小(即CTRGCN_joint.pdmodel)增 … See more 2024年10月12日最新提交所做的改进(当前V3):增加精度为91.157%的CTRCGN_lightV2,其GPU和CPU上推理速度对比分别比原模型分别快1.72倍和1.71倍! 2024年10 … See more 我们是要在NTU-RGB+D数据集、joint模态、X-sub评测标准下对我们的轻量化的CTRGCN模型进行训练、验证及推理。CTRGCN的数据准备详情见docs/zh-CN/dataset/ntu-rgbd.md,我们并没有完全按照其流程,即省略 … See more
GitHub - swsong89/ctrgcn
WebCTR-GCN/README.md Go to file Cannot retrieve contributors at this time 147 lines (103 sloc) 5.25 KB Raw Blame CTR-GCN This repo is the official implementation for Channel-wise Topology Refinement Graph … Webpyskl / configs / ctrgcn / ctrgcn_pyskl_ntu60_xview_hrnet / b.py Go to file Go to file T; Go to line L; Copy path Copy permalink; This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Cannot retrieve contributors at this time. 60 lines (57 sloc) 2.26 KB sohngyver.quickconnect.to
Issues · Uason-Chen/CTR-GCN · GitHub
WebJul 26, 2024 · Combining CTR-GC with temporal modeling modules, we develop a powerful graph convolutional network named CTR-GCN which notably outperforms state-of-the … WebCTR-GCN. This repo is the official implementation for Channel-wise Topology Refinement Graph Convolution for Skeleton-Based Action Recognition.The paper is accepted to ICCV2024. Note: We also provide a simple and strong baseline model, which achieves 83.7% on NTU120 CSub with joint modality only, to facilitate the development of skeleton … WebThis repo is the official implementation for ACFL, which is based on CTR-GCN repo. Introduction In this work, we are interested in the skeleton based action recognition with a focus on learning paradigm. Most existing methods tend to improve GCNs by leveraging multi-form skeletons due to their complementary cues. sohn gottes