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

Fix deepspeed default precision plugin amp_level to O2 #13897

Merged
merged 8 commits into from
Jul 29, 2022

Conversation

rohitgr7
Copy link
Contributor

@rohitgr7 rohitgr7 commented Jul 28, 2022

What does this PR do?

With

trainer = Trainer(
        strategy=DeepSpeedStrategy(zero_optimization=False),
        amp_backend='apex',
        gpus=2,
        precision=16,
        accelerator='gpu',
)

This error was raised:

RuntimeError: Unexpected optimization level None. Options are 'O0', 'O1', 'O2', 'O3'.  Note that in `O0`, `O1`, etc., the prefix O is the letter O, not the number zero.

But as per docs, default should be O2.

Does your PR introduce any breaking changes? If yes, please list them.

Before submitting

  • Was this discussed/approved via a GitHub issue? (not for typos and docs)
  • Did you read the contributor guideline, Pull Request section?
  • Did you make sure your PR does only one thing, instead of bundling different changes together?
  • Did you make sure to update the documentation with your changes? (if necessary)
  • Did you write any new necessary tests? (not for typos and docs)
  • Did you verify new and existing tests pass locally with your changes?
  • Did you list all the breaking changes introduced by this pull request?
  • Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

  • Is this pull request ready for review? (if not, please submit in draft mode)
  • Check that all items from Before submitting are resolved
  • Make sure the title is self-explanatory and the description concisely explains the PR
  • Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

cc @Borda @awaelchli @rohitgr7 @akihironitta @carmocca @justusschock

@rohitgr7 rohitgr7 added bug Something isn't working strategy: deepspeed precision: apex (removed) NVIDIA/apex precision labels Jul 28, 2022
@rohitgr7 rohitgr7 added this to the pl:1.6.x milestone Jul 28, 2022
@rohitgr7 rohitgr7 self-assigned this Jul 28, 2022
@github-actions github-actions bot added the pl Generic label for PyTorch Lightning package label Jul 28, 2022
@rohitgr7 rohitgr7 changed the title Fix deepspeed default precision plugin to O2 Fix deepspeed default precision plugin amp_level to O2 Jul 28, 2022
@rohitgr7 rohitgr7 marked this pull request as ready for review July 28, 2022 09:48
@rohitgr7 rohitgr7 requested a review from awaelchli July 28, 2022 15:29
@codecov
Copy link

codecov bot commented Jul 28, 2022

Codecov Report

Merging #13897 (468c37d) into master (aefb9ab) will increase coverage by 15%.
The diff coverage is 100%.

❗ Current head 468c37d differs from pull request most recent head eb3a02d. Consider uploading reports for the commit eb3a02d to get more accurate results

@@            Coverage Diff            @@
##           master   #13897     +/-   ##
=========================================
+ Coverage      61%      76%    +15%     
=========================================
  Files         335      333      -2     
  Lines       26301    26163    -138     
=========================================
+ Hits        16048    19980   +3932     
+ Misses      10253     6183   -4070     

@mergify mergify bot added the ready PRs ready to be merged label Jul 28, 2022
@rohitgr7 rohitgr7 enabled auto-merge (squash) July 29, 2022 08:56
@rohitgr7 rohitgr7 merged commit 0f6caff into master Jul 29, 2022
@rohitgr7 rohitgr7 deleted the fix/deepspeed_apx_lvl branch July 29, 2022 20:36
@awaelchli awaelchli modified the milestones: pl:1.6.x, pl:1.7 Aug 2, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working pl Generic label for PyTorch Lightning package precision: apex (removed) NVIDIA/apex precision ready PRs ready to be merged strategy: deepspeed
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants