simplify the KJT.split function when segment is the original KJT #3014
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
Summary:
# context
* in KJT.split function, when the segment == len(keys), the returned KJT contains the same data as the original KJT
* however in the function it recreates a new one which introduces extra cost
* this diff remove the redundent KJT creation
# analysis
* when segment == len(keys), start has to be zero so the stride_per_key_per_rank is the original one.
* the following KJT init produces the same KJT as self
```
KeyedJaggedTensor(
keys=self._keys,
values=self._values,
weights=self.weights_or_none(),
lengths=self._lengths,
offsets=self._offsets,
stride=self._stride,
stride_per_key_per_rank=stride_per_key_per_rank,
stride_per_key=None,
length_per_key=self._length_per_key,
lengths_offset_per_key=None,
offset_per_key=self._offset_per_key,
index_per_key=self._index_per_key,
jt_dict=self._jt_dict,
inverse_indices=None,
)
```
Differential Revision: D70756397
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D70756397 |
TroyGarden
added a commit
to TroyGarden/torchrec
that referenced
this pull request
May 29, 2025
…orch#3014) Summary: # context * in KJT.split function, when the segment == len(keys), the returned KJT contains the same data as the original KJT * however in the function it recreates a new one which introduces extra cost * this diff remove the redundent KJT creation # analysis * when segment == len(keys), start has to be zero so the stride_per_key_per_rank is the original one. * the following KJT init produces the same KJT as self ``` KeyedJaggedTensor( keys=self._keys, values=self._values, weights=self.weights_or_none(), lengths=self._lengths, offsets=self._offsets, stride=self._stride, stride_per_key_per_rank=stride_per_key_per_rank, stride_per_key=None, length_per_key=self._length_per_key, lengths_offset_per_key=None, offset_per_key=self._offset_per_key, index_per_key=self._index_per_key, jt_dict=self._jt_dict, inverse_indices=None, ) ``` Differential Revision: D70756397
TroyGarden
added a commit
to TroyGarden/torchrec
that referenced
this pull request
May 29, 2025
…orch#3014) Summary: # context * in KJT.split function, when the segment == len(keys), the returned KJT contains the same data as the original KJT * however in the function it recreates a new one which introduces extra cost * this diff remove the redundent KJT creation # analysis * when segment == len(keys), start has to be zero so the stride_per_key_per_rank is the original one. * the following KJT init produces the same KJT as self ``` KeyedJaggedTensor( keys=self._keys, values=self._values, weights=self.weights_or_none(), lengths=self._lengths, offsets=self._offsets, stride=self._stride, stride_per_key_per_rank=stride_per_key_per_rank, stride_per_key=None, length_per_key=self._length_per_key, lengths_offset_per_key=None, offset_per_key=self._offset_per_key, index_per_key=self._index_per_key, jt_dict=self._jt_dict, inverse_indices=None, ) ``` Differential Revision: D70756397
TroyGarden
added a commit
to TroyGarden/torchrec
that referenced
this pull request
May 29, 2025
…orch#3014) Summary: # context * in KJT.split function, when the segment == len(keys), the returned KJT contains the same data as the original KJT * however in the function it recreates a new one which introduces extra cost * this diff remove the redundent KJT creation # analysis * when segment == len(keys), start has to be zero so the stride_per_key_per_rank is the original one. * the following KJT init produces the same KJT as self ``` KeyedJaggedTensor( keys=self._keys, values=self._values, weights=self.weights_or_none(), lengths=self._lengths, offsets=self._offsets, stride=self._stride, stride_per_key_per_rank=stride_per_key_per_rank, stride_per_key=None, length_per_key=self._length_per_key, lengths_offset_per_key=None, offset_per_key=self._offset_per_key, index_per_key=self._index_per_key, jt_dict=self._jt_dict, inverse_indices=None, ) ``` Reviewed By: iamzainhuda Differential Revision: D70756397
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.
Summary:
context
analysis
Differential Revision: D70756397