目录

pct-jittor-modelnet40: Point Cloud Transformer for ModelNet40

This project is a Jittor implementation of the Point Cloud Transformer (PCT) for 3D shape classification on the ModelNet40 dataset. It was completed as part of the Graphics Experiment PA3.

Model Performance

  • Framework: Jittor
  • Target Accuracy: 80%
  • Achieved Accuracy: 87.64%
  • Hardware used for training: NVIDIA RTX 4090 (Cloud)

Requirements

  • Python >= 3.8
  • Jittor >= 1.3.1
  • NumPy

Running instructions

  1. Ensure the ModelNet40 dataset is located in the data/ directory.
  2. Run the script: ```bash python3 pct.py
关于

A Jittor implementation of Point Cloud Transformer (PCT) for ModelNet40 3D shape classification. This project is completed for Graphics Experiment PA3 with 87.64% accuracy

35.0 KB
邀请码
    Gitlink(确实开源)
  • 加入我们
  • 官网邮箱:gitlink@ccf.org.cn
  • QQ群
  • QQ群
  • 公众号
  • 公众号

版权所有:中国计算机学会技术支持:开源发展技术委员会
京ICP备13000930号-9 京公网安备 11010802032778号