I'm trying to figure out how to actually get an optimized model out of the pipeline in order to use as a gltf. It looks like the output of running the seurat pipeline with a capture from unity creates an obj and a png texture, but its both incredibly large (10mb-obj, 15mb png) and still looks pretty janky. However, loading it up in the viewer using the .ice file looks pretty good. So i'm curious, what is the proper way to get a 3d model file out that is good quality and small? The instructions don't really tell you much other than some flags you can use.

Is the .obj the final generated artifact i should be using?

Also, what is in the .ice file that lets it render properly?