PA3 - Point Cloud Transformer

A Jittor implementation of Point Cloud Transformer (PCT) for ModelNet40 classification.

Introduction

This project implements the Point Cloud Transformer (PCT) model using the Jittor deep learning framework for 3D point cloud classification on the ModelNet40 dataset.

Environment

• Python 3.12
• Jittor 1.3.11
• CUDA 12
• Google Colab

Dataset

ModelNet40 point cloud dataset.

Training data:

• train_points.npy
• train_labels.npy

Test data:

• test_points.npy

Usage

Train and generate prediction results:

python pct.py --epochs 50 --batch_size 32 --n_points 1024 --lr 0.01

Result

Accuracy: 0.8027