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

HF LLaMA -> megatron weight #10

Closed
dumpmemory opened this issue Aug 5, 2023 · 5 comments · Fixed by #27
Closed

HF LLaMA -> megatron weight #10

dumpmemory opened this issue Aug 5, 2023 · 5 comments · Fixed by #27
Assignees
@Olivia-fsm

Comments

@dumpmemory
Copy link
Contributor

Could you provide the script to convert hf llama or llama2 to megatron weight ? Thanks
@martinjaggi
Copy link
Contributor

there is a folder named after it:
https://github.com/epfLLM/Megatron-LLM/blob/main/weights2megatron/

@dumpmemory
Copy link
Contributor Author

I mean the checkpoint with huggingface transformers style to megatron weight. currently it just support meta weight style right ?

@martinjaggi
Copy link
Contributor

you're right. sorry we'll reopen the issue. BTW the megatron2HF converter (other direction) has just been improved and should work with no issue

@martinjaggi martinjaggi reopened this Aug 8, 2023
@Olivia-fsm Olivia-fsm self-assigned this Aug 11, 2023
@AleHD AleHD mentioned this issue Aug 13, 2023
Merged
5 tasks
@AleHD AleHD linked a pull request Aug 13, 2023 that will close this issue
Weight conversion testing and other features #27
Merged
5 tasks
@AleHD AleHD closed this as completed in #27 Aug 17, 2023
@dumpmemory
Copy link
Contributor Author

Thanks bro

@dumpmemory
Copy link
Contributor Author

dumpmemory commented Aug 20, 2023 •
edited
Loading

I found LLaMA-HF 1 hadn't been not supported from code #27 . Can we support it ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@Olivia-fsm Olivia-fsm

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Weight conversion testing and other features epfLLM/Megatron-LLM
3 participants
@martinjaggi @Olivia-fsm @dumpmemory