Adaptive Token Selection for Scalable Point Cloud Transformers

Welcome to the AdaPT Repository. This repository contains a deep learning model for point cloud classification using transformers and adaptive token dropping.

@article{baiocchi2025adapt,
title = {Adaptive token selection for scalable point cloud transformers},
journal = {Neural Networks},
year = {2025},
issn = {0893-6080},
doi = {https://doi.org/10.1016/j.neunet.2025.107477},
author = {Alessandro Baiocchi and Indro Spinelli and Alessandro Nicolosi and Simone Scardapane},
}

Installation

To get started with this project, follow these steps:

1. Clone the repository to your local machine:

   git clone https://github.com/ispamm/adaPT

2. Navigate to the project directory:

   cd adaPT

3. Create a Conda environment from the provided Adapt_env.yaml file:

   conda env create -f Adapt_env.yaml

4. Activate the Conda environment:

   conda activate Adapt_env

Usage

Once you have set up the Conda environment, you can run the code using the following steps:

1. Ensure you have activated the Conda environment as mentioned in the Installation section.

2. Run the main script:

   python main.py

   Replace main.py with the name of the main script in your project.

Hyperparameters

This project uses Hydra for hyperparameter management. Edit the config.yaml file to change the desired parameters.

Contributing

Contributions are welcome! If you find any issues or have suggestions for improvements, please open an issue or create a pull request. For major changes, please open an issue first to discuss the proposed changes.