Closed clumsy closed 8 months ago
Hi @rahul003, I was able to confirm this fix works by building and using a custom Docker image, I see:
Invoking script with the following command:
torchrun --nnodes 1 --nproc_per_node 1 -m <MY_MODULE> <MY_ARGS>
LGTM
@rahul003 Is there anything I can do help merging this PR?
Hi Alexander, The PR looks good to me, but I don’t have the authority to merge PRs here. We have to check with maintainers
Rahul Huilgol
On Wed, Jan 24, 2024 at 11:12 AM Alexander Jipa @.***> wrote:
@rahul003 https://github.com/rahul003 Is there anything I can do help merging this PR?
— Reply to this email directly, view it on GitHub https://github.com/aws/sagemaker-training-toolkit/pull/205#issuecomment-1908758617, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA2MBWHNJLAONELYXTF6PK3YQFMIFAVCNFSM6AAAAABAVF642GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYG42TQNRRG4 . You are receiving this because you were mentioned.Message ID: @.***>
Closing this one and open a new PR.
Issue #, if available:
Related to https://github.com/aws/sagemaker-python-sdk/pull/4324
Description of changes:
torchrun
can run modules via-m <MODULE
but SageMaker runtime does not yet support this entrypoint type. Adding a new entrypoint type and add support to torch_distributed.Testing done:
Merge Checklist
Put an
x
in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your pull request.General
Tests
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.