scale gradient accumulation steps with train batch size to keep effective batch size about the same #33281
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…tive batch size about the same
winglian
force-pushed
the
auto_find_batch_size_compensate
branch
from
ab90906 to
56cbbba
Compare
muellerzr
reviewed
Sep 3, 2024
Comment on lines
+1976
to
+1983
| def reduce_batch_size_fn_wrapper(train_batch_size, args): | ||
| effective_batch_size = args.gradient_accumulation_steps * train_batch_size | ||
| def reduce_batch_size_fn(): | ||
| nonlocal train_batch_size | ||
| train_batch_size = train_batch_size // 2 | ||
| args.gradient_accumulation_steps = effective_batch_size // train_batch_size | ||
| return train_batch_size | ||
| return reduce_batch_size_fn |
There was a problem hiding this comment.
Let's put this in trainer_utils I think since that's where we have find_executable_batch_size defined.
| args.gradient_accumulation_steps = effective_batch_size // train_batch_size | ||
| return train_batch_size | ||
| return reduce_batch_size_fn | ||
|
|
||
| inner_training_loop = find_executable_batch_size( |
There was a problem hiding this comment.
The trainer_utils's find_executable_batch_size will then also need changing.
Also need to be careful about versioning inside that func
There was a problem hiding this comment.
It should be set as >=0.34.0
There was a problem hiding this comment.
sorry, I forgot to commit a file. should be there now
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
goes with huggingface/accelerate#3071
Fixes # (issue)
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
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.