Allow to warm up a dataset to avoid re-reading (validation) batches#40
Merged
ravwojdyla merged 1 commit intomainfrom Jul 16, 2025
Merged
Allow to warm up a dataset to avoid re-reading (validation) batches#40ravwojdyla merged 1 commit intomainfrom
ravwojdyla merged 1 commit intomainfrom
Conversation
ravwojdyla
force-pushed
the
rav-skip-prob-batch
branch
from
4158d42 to
3848616
Compare
ravwojdyla
force-pushed
the
rav-prefill-val-ds
branch
from
0d599fb to
8b2e4cc
Compare
ravwojdyla
force-pushed
the
rav-prefill-val-ds
branch
from
8b2e4cc to
5e69ba6
Compare
yonromai
approved these changes
Jul 16, 2025
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
In our experiments we often run validation on subset of the validation dataset (e.g. 32 batches) relatively frequently (e.g. every 32 steps), as well as longer validation at the end of training. The way we have this implemented right now in lightning will reread the validation dataset on each validation, this creates a gap in the GPU utilization:
To avoid that we can keep the intermediate validation batches in memory.