Pipeline: Add support to eval micro bs configuration #4859
Merged
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
nelyahu
requested review from
ShadenSmith,
duli2012,
mrwyattii and
tjruwase
as code owners
loadams
reviewed
Jan 2, 2024
When running evaluation the general memory consumption is reduced. Mainly due to absence of gradients, and hanging FWD activations. It allows to increase the micro-bs and improve the evaluation performance. This commits add the option to pass num_micro_batches to eval_batch(), as the current assumption is that same micro-bs and global-bs is used, so same number micro batches will take place. This commit also modifies _scale_loss_by_gas in runtime/engine.py to consider number of eval micro batches for loss scaling instead of training gas.
nelyahu
force-pushed
the
pp_eval_micro_bs
branch
from
18ed89d to
6360e29
Compare
loadams
approved these changes
Jan 3, 2024
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
mauryaavinash95
pushed a commit
to mauryaavinash95/DeepSpeed
that referenced
this pull request
Feb 17, 2024
When running evaluation the general memory consumption is reduced. Mainly due to absence of gradients, and hanging FWD activations. It allows to increase the micro-bs and improve the evaluation performance. This commits add the option to pass num_micro_batches to eval_batch(), as the current assumption is that same micro-bs and global-bs is used, so same number micro batches will take place. This commit also modifies _scale_loss_by_gas in runtime/engine.py to consider number of eval micro batches for loss scaling instead of training gas. Co-authored-by: Logan Adams <[email protected]>
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.
When running evaluation the general memory consumption is reduced. Mainly due to absence of gradients, and hanging FWD activations. It allows to increase the micro-bs and improve the evaluation performance. This commits add the option to pass num_micro_batches to eval_batch(), as the current assumption is that same micro-bs and global-bs is used, so same number micro batches will take place.
This commit also modifies _scale_loss_by_gas in runtime/engine.py to consider number of eval micro batches for loss scaling instead of training gas.