Implement __torch_function__ for KeyedTensor #3329
Open
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
meta-cla
bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D81047278 |
jd7-tr
added a commit
to jd7-tr/torchrec
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
force-pushed
the
export-D81047278
branch
from
0823584 to
c9bac80
Compare
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
|
This pull request was exported from Phabricator. Differential Revision: D81047278 |
jd7-tr
added a commit
to jd7-tr/torchrec
that referenced
this pull request
Aug 28, 2025
Summary: Pull Request resolved: pytorch#3329 X-link: pytorch/pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
force-pushed
the
export-D81047278
branch
from
c9bac80 to
b9821f1
Compare
|
This pull request was exported from Phabricator. Differential Revision: D81047278 |
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 Pull Request resolved: pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
added a commit
to jd7-tr/torchrec
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
force-pushed
the
export-D81047278
branch
from
b9821f1 to
87c1e8c
Compare
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
Summary: Pull Request resolved: pytorch#3329 X-link: pytorch/pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Reviewed By: malaybag Differential Revision: D81047278
|
This pull request was exported from Phabricator. Differential Revision: D81047278 |
jd7-tr
force-pushed
the
export-D81047278
branch
from
87c1e8c to
ab9d50c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D81047278 |
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 Pull Request resolved: pytorch#161683 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
jd7-tr
added a commit
to jd7-tr/pytorch
that referenced
this pull request
Aug 28, 2025
Summary: X-link: pytorch/torchrec#3329 1. There are a bunch of `torch.ops.aten` operations that can't handle `KeyedTensor`: The error was occurring because these ops expects a regular `Tensor` but was receiving a `KeyedTensor` object. 2. Implement `__torch_function__` for `KeyedTensor`, so when these incompatible operations are called with a `KeyedTensor`, the `__torch_function__` method automatically delegates the op to the underlying values tensor from the `KeyedTensor` and returns a new `KeyedTensor` with updated values. Test Plan: ``` buck2 run mode/opt fbcode//aps_models/ads/gmp:launcher_with_publish mode=mtml_mobile_cvr_model/managed/Y2025Q2/local_mode_mtml_mobile_cvr_model_733415799_v0_fork +training.ir_serializer=manifold ``` MAST job: https://fburl.com/mlhub/pp937uxf Rollback Plan: Reviewed By: malaybag Differential Revision: D81047278
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:
X-link: pytorch/pytorch#161683
There are a bunch of
torch.ops.atenoperations that can't handleKeyedTensor: The error was occurring because these ops expects a regularTensorbut was receiving aKeyedTensorobject.Implement
__torch_function__forKeyedTensor, so when these incompatible operations are called with aKeyedTensor, the__torch_function__method automatically delegates the op to the underlying values tensor from theKeyedTensorand returns a newKeyedTensorwith updated values.Reviewed By: malaybag
Differential Revision: D81047278