Implement named tensor unflatten(dim, namedshape).#25658
Closed
zou3519 wants to merge 5 commits intogh/zou3519/138/basefrom
Closed
Implement named tensor unflatten(dim, namedshape).#25658zou3519 wants to merge 5 commits intogh/zou3519/138/basefrom
unflatten(dim, namedshape).#25658zou3519 wants to merge 5 commits intogh/zou3519/138/basefrom
Conversation
zou3519
added a commit
that referenced
this pull request
Sep 4, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 4d552f0 Pull Request resolved: #25658
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: [D17192655](https://our.internmc.facebook.com/intern/diff/D17192655)
zou3519
added a commit
that referenced
this pull request
Sep 6, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 5be58c6 Pull Request resolved: #25658
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: [D17192655](https://our.internmc.facebook.com/intern/diff/D17192655)
zou3519
added a commit
that referenced
this pull request
Sep 9, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 3207bd8 Pull Request resolved: #25658
nairbv
reviewed
Sep 9, 2019
| - func: align_tensors(Tensor[] tensors) -> Tensor[] | ||
| named_guard: False | ||
|
|
||
| - func: unflatten(Tensor self, Dimname dim, int[] sizes, DimnameList names) -> Tensor |
Collaborator
There was a problem hiding this comment.
is there any close equivalent in numpy?
Contributor
Author
There was a problem hiding this comment.
No, in numpy one would use reshape. I did some research into the naming of this a while back, below is a summary:
In xarray (which is numpy + named axis), they use stack and unstack. The naming is based on pandas arrays' stack and unstack, which work in a similar manner.
In addition, Sasha Rush's namedtensor design uses stack and split.
Finally, salesforce/matchbox (autobatching with pytorch tensors) uses split_dim.
My proposal is to use flatten and unflatten instead of the above names. This is because stack and split already have semantic meaning: in pytorch, stack is an operation that takes a list of tensors and returns a new tensor; split is an operation that takes a tensor and splits it into multiple other tensors. Something like stack_dim and split_dim would make the semantics clearer, but since we already have flatten as a part of the API, unflatten is a reasonable counterpart.
Collaborator
There was a problem hiding this comment.
@zou3519 actually this probably also needs an entry in the docs
Contributor
Author
There was a problem hiding this comment.
No named tensors docs yet, but they'll come before the release.
nairbv
approved these changes
Sep 9, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: [D17192655](https://our.internmc.facebook.com/intern/diff/D17192655)
zou3519
added a commit
that referenced
this pull request
Sep 10, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 97f0895 Pull Request resolved: #25658
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: [D17192655](https://our.internmc.facebook.com/intern/diff/D17192655)
zou3519
added a commit
that referenced
this pull request
Sep 11, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 492d90f Pull Request resolved: #25658
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: [D17192655](https://our.internmc.facebook.com/intern/diff/D17192655)
zou3519
added a commit
that referenced
this pull request
Sep 17, 2019
This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] ghstack-source-id: 74c497c Pull Request resolved: #25658
zdevito
pushed a commit
to zdevito/ATen
that referenced
this pull request
Sep 18, 2019
Summary: Pull Request resolved: pytorch/pytorch#25658 This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: D17192655 Pulled By: zou3519 fbshipit-source-id: fd9bd2f462c23a4df1c23d66f2aa95076ff1b160
Contributor
laurentdupin
pushed a commit
to laurentdupin/pytorch
that referenced
this pull request
Apr 24, 2026
Summary: Pull Request resolved: pytorch#25658 This unflattens `dim` according to the shape specified in `namedshape`. `namedshape` may be either an OrderedDict or an iterable of (name, size) tuples. Future: - It is possible to make it take a dict in Python >= 3.6 because those are ordered by default, but I'll leave that task for the future. Test Plan: - new tests [namedtensor ci] Differential Revision: D17192655 Pulled By: zou3519 fbshipit-source-id: fd9bd2f462c23a4df1c23d66f2aa95076ff1b160
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
5 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.
Stack from ghstack:
unflatten(dim, namedshape). #25658 Implement named tensorunflatten(dim, namedshape).This unflattens
dimaccording to the shape specified innamedshape.namedshapemay be either an OrderedDict or an iterable of (name, size)tuples.
Future:
ordered by default, but I'll leave that task for the future.
Test Plan:
Differential Revision: D17192655