-
Notifications
You must be signed in to change notification settings - Fork 4.2k
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
Failing to get proper visual observation on a Linux build using HDRP #3835
Comments
Hi, What are the graphics settings you have on Linux ? (Are you using Vulkan ?) |
Hey! Yes, I'm using Vulkan, on a GTX 1060 GPU. Yes, I can run the environment with zero issues on Linux. I tried to run it just by calling The Player logs was a good call! Here you go:
Thanks in advance Vincent! |
So it seems related with occlusion culling? |
OK I disabled ambient occlusion just in case and all post processing. Run it one more time and got on logs:
While searching about that new error |
Maybe this is due to our use |
Great catch Vincent! I commented that out and now it works perfectly: The error message related with culling is still there but no more complains about missing textures. Inspired by your suggestion I tried to change settings to allow resizable window like: However that didn't work neither. I put all the post processing back and your solution works. Thank you! |
Should I open a PR for that? |
Hi @automata, |
@surfnerd , We could do # if not PLATFORM_LINUX and not EDITOR and VERSION_2019_3
Screen.SetResolution(width, height)
#endif to make sure we do not set the resolution on this platform. |
…This is to help relove #3835 since setting the screen resolution on Linux 2019.3 can cause issues.
Tracking this bug internally : MLA-961 |
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
Describe the bug
While using the HDRP pipeline, we're able to create and train following the steps from https://github.com/Unity-Technologies/ml-agents/blob/latest_release/docs/Learning-Environment-Create-New.md.
We can even run a Jupyter notebook, send actions and get both vector and visual observations from the running environment inside Unity Editor.
However, when we build for Linux and try to call the binary from the Jupyter notebook, we're able to run actions but resulting visual observation is the right camera shot the first time but just dark images (with some incomplete representations as next results), as you can see in the following screenshot.
We tested the same project on Windows 10 and it works, from inside Unity Editor or a built binary for Windows.
To Reproduce
Steps to reproduce the behavior:
Console logs / stack traces
We don't see any error or other relevant information on console logs.
Screenshots
![Screenshot from 2020-04-23 17-33-32](https://user-images.githubusercontent.com/49062/80147305-b2807c00-8589-11ea-950c-263e1fd63b55.png)
Environment:
The text was updated successfully, but these errors were encountered: