Skip to content

Command line interface

DeePoints comes with a powerful set of builtin commands to manipulate point cloud datasets and to train a deep learning archicture on such data.

Note

All the commands explained in this section must be run within the installed virtual environment, so do not forget to run the following command beforehand.

source .venv/bin/activate # or source .venv/bin/activate.fish if you use fish

python .\venv\Scripts\activate

Dealing with point clouds

The script file located at scripts/preprocessing.py allows for easy manipulation of entire directories full of point cloud files of multiple formats, including .parquet and .laz. Deep dive into the functionalities at this page. You will learn more about converting point clouds across different file formats and splitting them in smaller chunks for better deep learning processing.

Building a deep learning model

You can also spawn the training of a deep learning model of your choice using the script located at scripts/segmentation.py. Learn here how to pass different parameters to customize the training of a point cloud segmentation/classification model the way you like.