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

JAX as benchmark #32

Open
pedronahum opened this issue Dec 28, 2024 · 4 comments
Open

JAX as benchmark #32

pedronahum opened this issue Dec 28, 2024 · 4 comments

Comments

@pedronahum
Copy link

Hi all,

Wondering if you would be interested in benchmarking the performance of Swift vs JAX (similar to what you have done with Pytorch and Tensorflow).

Initial performance tests suggest JAX is much faster than Pytorch and TensorFlow, but still slower than Swift. Prelim results here: https://gist.github.com/pedronahum/61f9c96bcea6d6251a4406162d0fcdc1.

Best,

Pedro N.

@JaapWijnen
Copy link
Collaborator

Hi Pedro, definitely interested! I'll take a look after the holiday break!

@pedronahum
Copy link
Author

Perfect! In the meantime, I will try to update the gist to include a TF version that leverages XLA. The message appears to be same, nonetheless, Swift looks to be faster than the AI/ML battle-tested XLA compiler in calculating gradients. But I let you do the official performance measurements :). Enjoy the holiday break.

@JaapWijnen
Copy link
Collaborator

Would you be interested in making a pull request?

@pedronahum
Copy link
Author

Sure, give me a couple of days to move the code from colab and add a relevant folder with instructions.

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