Add pinned_memory and non_blocking transfer for default collate_fn #52948
Conversation
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
Signed-off-by: Srinath Krishnamachari <[email protected]>
Signed-off-by: Srinath Krishnamachari <[email protected]>
Signed-off-by: Srinath Krishnamachari <[email protected]>
Signed-off-by: Srinath Krishnamachari <[email protected]>
Signed-off-by: Srinath Krishnamachari <[email protected]>
Co-authored-by: lanbochen-anyscale <[email protected]> Signed-off-by: srinathk10 <[email protected]>
…athk10-train-fix-collate-fn
Signed-off-by: Srinath Krishnamachari <[email protected]>
Signed-off-by: Srinath Krishnamachari <[email protected]>
srinathk10
requested review from
hongpeng-guo,
justinvyu,
matthewdeng,
raulchen,
woshiyyya and
a team
as code owners
srinathk10
changed the base branch from
master
to
srinathk10-train-fix-collate-fn
Signed-off-by: srinathk10 <[email protected]>
…-memory Signed-off-by: srinathk10 <[email protected]>
srinathk10
changed the base branch from
srinathk10-train-fix-collate-fn
to
master
Signed-off-by: srinathk10 <[email protected]>
raulchen
reviewed
May 13, 2025
| @@ -213,6 +213,13 @@ def __call__(self, batch: "pyarrow.Table") -> Dict[str, List["torch.Tensor"]]: | |||
| # However, for CPU transfer, we need to combine the chunked arrays first | |||
| # before converting to numpy format and then to Tensors. | |||
| combine_chunks = self.device.type == "cpu" | |||
|
|
|||
| # If the device is CPU, we don't need to pin the memory. | |||
| pin_memory = self.device.type != "cpu" | |||
There was a problem hiding this comment.
We should probably expose this arg to users.
pinning memory isn't always better (e.g., when there are many small batches)
There was a problem hiding this comment.
yea, seeing overhead with pinning for batch size = 32
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.
Why are these changes needed?
Add pinned_memory and non_blocking transfer for default collate_fn
Related issue number
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/under thecorresponding
.rstfile.