Skip to content

Check dist.is_initialized before calling dist.get_rank()#194

Open
thanh-lam wants to merge 1 commit intofoundation-model-stack:mainfrom
thanh-lam:thanh-lam/save-val-1aiu
Open

Check dist.is_initialized before calling dist.get_rank()#194
thanh-lam wants to merge 1 commit intofoundation-model-stack:mainfrom
thanh-lam:thanh-lam/save-val-1aiu

Conversation

@thanh-lam
Copy link
Copy Markdown

@thanh-lam thanh-lam commented Mar 20, 2026

This PR fixes the aiu-apps-sw-tracker issue #1688.

When --save_validation_info_outputs is requested, in case of single aiu, calling dist.get_rank() will fail. Therefore, checking dist.is_initialized() is required. Then, just simply set the bool=1.

Signed-off-by: Thanh Lam <thanhlam@us.ibm.com>
@thanh-lam thanh-lam force-pushed the thanh-lam/save-val-1aiu branch from 57fd859 to 9d612ad Compare March 20, 2026 18:43
@thanh-lam
Copy link
Copy Markdown
Author

Hi @gkumbhat - Thanks for helping create this PR!

Did you or someone in your team review it? It's very simple change on one line and doesn't have any other dependencies. Let me know and I can invite others by comment.

I don't have access to this repo to assign reviewers.

@JamesMartine
Copy link
Copy Markdown

@matthew-pisano This PR may help with getting DPP running on single AIU

Copy link
Copy Markdown
Contributor

@gkumbhat gkumbhat left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me!

@thanh-lam
Copy link
Copy Markdown
Author

@gkumbhat Thanks for approving! I see there're still "2 workflows awaiting approval". Who can we ask to help with reviewing? @Nhan-Hoang Can you help?

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

Successfully merging this pull request may close these issues.

3 participants