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

Speeding up training stage #130

Open
hardikdava opened this issue Sep 28, 2023 · 4 comments
Open

Speeding up training stage #130

hardikdava opened this issue Sep 28, 2023 · 4 comments

Comments

@hardikdava
Copy link

hardikdava commented Sep 28, 2023

I am new to gaussian splatting and taichi as well. Is there quick way to speed up training process without loosing quality. Is it possible to do batch processing?

P.S. The training process takes more than 1 hour for 20k iterations on Tesla T4.

@wanmeihuali
Copy link
Owner

I think so. It's in previous TODO list. But I don't have that much time to implement it... It shall not be difficult, just need to add a new dim in sorting key(tile uv to duv), and also add the dim in rasterization/backward. As we only use less than 4GB GPU memory now, I think it shall help.

@jb-ye
Copy link
Contributor

jb-ye commented Nov 16, 2023

@wanmeihuali do you still have plans for batch processing?

@wanmeihuali
Copy link
Owner

@Alexma3312 @Alexma3312 @yanzhoupan Do you have time to take a look?

@yanzhoupan
Copy link
Collaborator

Will start working on batch training.

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

4 participants