sagemaker-training-toolkit icon indicating copy to clipboard operation
sagemaker-training-toolkit copied to clipboard

Published 20 hours ago • verified publisher icon aws

feature: Add torch_distributed support for Trainium instances in SageMaker

Open satishpasumarthi opened this issue 2 years ago • 10 comments

Issue #, if available:

Description of changes:

  • Add support for torch_distributed (torchrun) distribution strategy for trainium instances in SageMaker.
  • PySDK PR: https://github.com/aws/sagemaker-python-sdk/pull/3424
  • PT Toolkit PR: https://github.com/aws/sagemaker-pytorch-training-toolkit/pull/248
  • CONTRIBUTING.md: Updated the python version to py37 since our package doesn't work on py36 anymore because of asyncio requirement

Testing done:

  • Added unit tests for single node and multi node scenarios

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

  • [x] I have read the CONTRIBUTING doc
  • [x] I used the commit message format described in CONTRIBUTING
  • [ ] I have used the regional endpoint when creating S3 and/or STS clients (if appropriate)
  • [x] I have updated any necessary documentation, including READMEs

Tests

  • [x] I have added tests that prove my fix is effective or that my feature works (if appropriate)
  • [x] I have checked that my tests are not configured for a specific region or account (if appropriate)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

satishpasumarthi avatar Oct 14 '22 05:10 satishpasumarthi

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: ed43264559fe3fe1ba5c9d0d11d0a8a6cd7b1d33
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 14 '22 05:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 029bcf91858c77c787accf20edda7b8469275f9b
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 14 '22 05:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: fc9dabae3ca449c0753543e39f426d2e3d93d026
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 16 '22 18:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: d0a645445ada8b3f91bdc85d7801b55304c6b400
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 16 '22 22:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 12acccb8f58036bcf7570955541107e21bfc971c
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 16 '22 23:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: ab67c7e674cab0b1c45fa7746cafaba6ece547f2
  • Result: SUCCEEDED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 17 '22 06:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: f20e27e3d2c83f308e336b3f66575897f8496a70
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 17 '22 20:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 0314b9f4886976a388df511eaaef5fb3660f2c3d
  • Result: SUCCEEDED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 17 '22 22:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: eb5097d64f22fcb6341cd73cfb59f80212326f31
  • Result: SUCCEEDED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 20 '22 05:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 2aa23ea7e1cff28e4dfdc291eae5237d62f5dd1e
  • Result: FAILED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 20 '22 07:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 384363f1138413b42a5a8325637b7d0bdff6d478
  • Result: SUCCEEDED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 20 '22 15:10 sagemaker-bot

AWS CodeBuild CI Report

  • CodeBuild project: sagemaker-training-toolkit-pr
  • Commit ID: 32e999817dd15c1054d190b7478c268c02333648
  • Result: SUCCEEDED
  • Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

sagemaker-bot avatar Oct 20 '22 17:10 sagemaker-bot