Open Zenif-NIght opened 1 year ago
Hi there, it is expected that the two files are of the same size as the original file. The utility script is only updating the references of the meshes of the original asset and marking certain objects in the file as instanceable. This will be processed by the USD framework when loaded in Isaac Sim, which will reduce the memory consumption required for the asset.
Issue: Problem with converting USD files to Instanceable Assets
Description
I am encountering an issue while attempting to convert normal USD files to Instanceable USD Assets. I have followed the provided guide and tried using the URDF file and the converter, but I have been unsuccessful in achieving the desired results.
Background
My first attempt at converting the files was to use the URDF file and the provided converter. However, the resulting
go1.py
andgo1.usd
files, which can be found here and here, respectively, seemed to be fixed in space in some way. I also tried a different approach using the larger NVIDIA USD forgo1
file from this location it works, but is too large to work with. I need to use Instanceable USD Assets for my project.I have reached out on Discord and asked a question regarding this issue, which you can find here.
Issue Details
I am new to creating USD files, and I attempted to create an Instanceable
Go1
asset following the guide provided here. Unfortunately, I did not achieve the desired outcome.I downloaded the original
go1.usd
file from Isaac Sim, and after running the code provided in thecreate_instanceable_assets.py
script, I obtainedgo1_instanceable.usd
andgo1_instanceable_meshes.usd
files. However, both of these files have the same size as the originalGo1.usd
file, which indicates that the conversion process did not work as intended.To convert the asset to an Instanceable Asset, I ran the following code in the Script Editor:
Steps to Reproduce
Download the necessary files and dependencies:
Additional Information
Ubuntu 20.04.6 LTS, CPU: AMD Ryzen 9 5900X 12-Core Processor, Mem: 62GB, NVIDIA GeForce RTX 3090