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

How to load codegen models in jaxformer locally?? #13

Open
harry-stark opened this issue Nov 23, 2022 · 1 comment
Open

How to load codegen models in jaxformer locally?? #13

harry-stark opened this issue Nov 23, 2022 · 1 comment

Comments

@harry-stark
Copy link

harry-stark commented Nov 23, 2022

I've tried to extended the local master with load_ckpt function, but ran into trouble with:

assert jax.process_count() == ckpt['process_count']

Any pointers for this?? Do we've to use jax.distributed.initialize or is there another way for doing it with a single process ??

Edit: Conversion script for jax to hf model? Does one can adopt gptj script of conversion??

Setup: single node with 8 A100s

cc @enijkamp

@glicerico
Copy link

hey @harry-stark , I am having problem with the same assertion, even running in gcloud TPUs, did you solve this?

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