Trying to run colmap for NeRF pipeline on Jetson Xavier NX using instant-ngp

Dev Environment:
GPU: Jetson Xavier NX
Architecture: aarch64
Linux OS: Ubuntu 20.04

I’ve managed to build instant-ngp on the aarch64 architecture and ran sample datasets, provided by NeRF Synthetic and instant-ngp, to test whether the Jetson Xavier NX could render the NeRF. It was successful in running the sample datasets and objects, but I wanted to create a way, using the same hardware specs, to run my own datasets, from video input provided by me to mesh output. The problem is, the formatted data that instant-ngp accepts requires colmap, a program I am having trouble running on Jetson since it doesn’t seem to be compatible with aarch64 architecture (Note: ran Installation — COLMAP 3.9-dev documentation installation on Ubuntu 20.04 x86_64 system and had 0 issues). If colmap doesn’t work, I’m looking for alternatives that provide the same functionality for extracting features from an image dataset and/or render NeRFs on Jetson Xavier NX hardware and aarch64 architecture. Unless this is purely an issue of the architecture not being supported for this kind of task. Either way, I’m curious to see if anyone has gotten around this issue.