Skip to content

Multiply lr scheduler steps by num_processes.#3983

Merged
sayakpaul merged 2 commits into
huggingface:mainfrom
eliphatfs:main
Jul 13, 2023
Merged

Multiply lr scheduler steps by num_processes.#3983
sayakpaul merged 2 commits into
huggingface:mainfrom
eliphatfs:main

Conversation

@eliphatfs

Copy link
Copy Markdown
Contributor

What does this PR do?

Fixes #3954.

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@sayakpaul @muellerzr

@HuggingFaceDocBuilderDev

HuggingFaceDocBuilderDev commented Jul 7, 2023

Copy link
Copy Markdown

The documentation is not available anymore as the PR was closed or merged.

@sayakpaul sayakpaul left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Thanks much!

@eliphatfs

Copy link
Copy Markdown
Contributor Author

Is this supposed to be merged now?

@sayakpaul

Copy link
Copy Markdown
Member

Oh I was waiting for @muellerzr's eyes on this too.

@eliphatfs

Copy link
Copy Markdown
Contributor Author

Please do not merge yet. It seems that we should not multiply the steps by accumulation.

@sayakpaul sayakpaul requested a review from muellerzr July 13, 2023 12:11

@muellerzr muellerzr left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Thanks!

@sayakpaul sayakpaul merged commit ece5522 into huggingface:main Jul 13, 2023
orpatashnik pushed a commit to orpatashnik/diffusers that referenced this pull request Aug 1, 2023
* Multiply lr scheduler steps by `num_processes`.

* Stop multiplying steps by gradient accumulation.
orpatashnik pushed a commit to orpatashnik/diffusers that referenced this pull request Aug 1, 2023
* Multiply lr scheduler steps by `num_processes`.

* Stop multiplying steps by gradient accumulation.
orpatashnik pushed a commit to orpatashnik/diffusers that referenced this pull request Aug 1, 2023
* Multiply lr scheduler steps by `num_processes`.

* Stop multiplying steps by gradient accumulation.
AmericanPresidentJimmyCarter pushed a commit to AmericanPresidentJimmyCarter/diffusers that referenced this pull request Apr 26, 2024
* Multiply lr scheduler steps by `num_processes`.

* Stop multiplying steps by gradient accumulation.
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.

Wrong learning rate scheduler training step count for examples with multi-gpu.

4 participants