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.

Sign up for GitHub

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

Jump to bottom

Does this implementation allow serving a single model to multiple devices (cuda/tpu/cpu?) #3

Open
CompRhys opened this issue Nov 17, 2023 · 1 comment
Open

Does this implementation allow serving a single model to multiple devices (cuda/tpu/cpu?) #3

CompRhys opened this issue Nov 17, 2023 · 1 comment

Comments

@CompRhys
Copy link

There are some issues with using a GPU trained model on CPU for the main package (see: ACEsuit/mace#153). Do the same issues arise when trying to swap backend with the JAX version?
Have you done any testing of the inference speed up of the CPU version here?
@ilyes319
Copy link
Contributor

Hey @CompRhys,

I think CPU is faster on JAX compared to torch below 2.0. It should be quite easy to convert a model to CPU.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@CompRhys @ilyes319