Running Inferences on Real-Time Data

[JordySehn]

Hi, I am trying to adopt your network to do real-time lidar segmentation on a 64 beam ouster lidar (as I have done in the past on other semantic kitti trained networks) but seem to be having some issues getting the network to run a test inference when you just have the point cloud alone in a .bin format.

The TLDR: How can I run inferences on the semantic KITTI dataset using only the point clouds and not the image pairs as you say you can in your paper?