Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Models take a long time to import into Unity #313

Open
elephantpanda opened this issue Feb 28, 2023 · 1 comment
Open

Models take a long time to import into Unity #313

elephantpanda opened this issue Feb 28, 2023 · 1 comment

Comments

@elephantpanda
Copy link

elephantpanda commented Feb 28, 2023

When adding an onnx asset to Unity, it takes a very long time for Baracuda to preprocess it. This is because it reads the onnx and writes out all the inputs, layer names etc.

For a large onnx like 2GB this can take several minutes.

Is this really necessary? There should be a way to import the onnx file without all the preprocessing, or at least to delay this until the user wants to use the onnx file.

I just want to be able to drag an Onnx file into the Assets folder without it taking 5 minutes.

Perhaps the developers could come up with a solution to this problem?

(Also is it doing this preprocessing on the CPU or GPU? It should be GPU-accelerated if possible!)

Actually, if you are doing this on a computer without a graphics card, you probably don't want it to do any preprocessing. But later when you open the project on a computer with a GPU you do. - Or maybe you don't since loading the onnx onto the GPU could take a long time.

@cullenwire1
Copy link

Hey @pauldog I noticed you are pretty active on this repo. If you would like to talk to Unity directly about this please grab a time w/ me here, or email me at bill dot cullen at unity3d dot com. I am seeking feedback about Barracuda to inform our future plans.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants