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

Question on the estimated memory of GaLore #67

Open
zqOuO opened this issue Dec 14, 2024 · 0 comments
Open

Question on the estimated memory of GaLore #67

zqOuO opened this issue Dec 14, 2024 · 0 comments

Comments

@zqOuO
Copy link

zqOuO commented Dec 14, 2024

Thank you for your great work.
I am trying to reproduce the memory results in Table 2: Comparison with low-rank algorithms on pre-training various sizes of LLaMA models on C4 dataset.
In your paper, the estimated memory for LLaMa 350M is 1.22G. I conducted experiment for 350M and got the following:
Total Params: 367.97M
GaLore enabled: 302.38M
Is there an exact equation for estimating the memory usage for GaLore for different models?

@zqOuO zqOuO changed the title Question on reproducing the estimated memory of GaLore Question on the estimated memory of GaLore Dec 14, 2024
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

1 participant