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

colmap stage 0 failed #5

Open
r530044129 opened this issue Mar 9, 2023 · 4 comments
Open

colmap stage 0 failed #5

r530044129 opened this issue Mar 9, 2023 · 4 comments

Comments

@r530044129
Copy link

image
image

I just follow your tutorial, which use garden to generate, i not sure why i failed

the stage 0 of nerf is success

@r530044129
Copy link
Author

I use ubuntu 18, python 3.8, torch 1.13.1, cuda 11.6

@ashawkey
Copy link
Owner

ashawkey commented Mar 9, 2023

@r530044129 Hi, have you modified the code? NVdiffrast is not supposed to run in CUDA context in non-gui mode as there are many limitations, like this one.
You could make H/W divisible by 8 before passing it to ``mark_unseen_triangles` if you have to use CUDA context, but it will still error out in texture exportation (unless you reduce texture resolution to 2048 and ssaa to 1, since the max resolution CUDA context support is 2048).

@r530044129
Copy link
Author

@ashawkey yeah, I forget tell you that I do have changed code, in mark_unseen_triangles(),I change OpenGL to Cuda, around line 818.
Because When I use default code, there is an Aborted after Saving Mesh, see this picture
[F glutil.cpp:338] eglInitialize() failed
image
And after some google search, someone suggest maybe I could use cuda to replace Opengl, but it's seems not work too, lol
Hope you can help me to figure out the Aborted Problem, thanks!

@ashawkey
Copy link
Owner

This indicates the OpenGL installation is somewhat wrong. You may refer to this sicxu/Deep3DFaceRecon_pytorch#2.

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