This repository contains the notebook and model for fast point cloud segmentation with model predictions. The notebook shows how you can set up model-assisted labeling on Segments.ai using its simple Python SDK. As an example, a set of diverse frames of the SemanticKITTI dataset are labeled by using model predictions from SqueezeSegV3.
A write-up of this notebook with additional information on how to set up model-assisted labeling for cuboid annotation can be found at segments.ai/blog/fast-point-cloud-labeling-with-model-predictions