Make logsumexp_out inplace #9755
Closed
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
t-vi
requested review from
apaszke,
colesbury,
ezyang,
gchanan,
soumith and
zdevito
as code owners
zou3519
reviewed
Jul 24, 2018
| b = torch.zeros(5, 2) | ||
| c = b[:, 0] | ||
| torch.logsumexp(a, 1, out=c) | ||
| self.assertTrue(np.allclose(expected, b[:, 0].numpy())) |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
Contributor
|
Sorry did not mean to delete that comment. For context, I was asking why the |
zou3519
reviewed
Jul 24, 2018
aten/src/ATen/native/ReduceOps.cpp
Outdated
| AT_CHECK(self.numel() != 0, "logsumexp only works on nonempty tensors"); | ||
| auto maxes = at::max_values(self, dim, true); | ||
| auto maxes_squeezed = (keepdim ? maxes : maxes.squeeze(dim)); | ||
| maxes_squeezed.masked_scatter_(maxes_squeezed.abs() == INFINITY, at::zeros({1}, maxes_squeezed.options()).expand_as(maxes_squeezed)); |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
zou3519
approved these changes
Jul 24, 2018
Contributor
zou3519
left a comment
zou3519
left a comment
There was a problem hiding this comment.
LGTM, would prefer at::zeros({1}, maxes_squeezed.options()).expand_as(maxes_squeezed) written as at::zeros_like(maxes_squeezed) if that works
zou3519
approved these changes
Jul 24, 2018
Contributor
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@zou3519 has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
Contributor
|
@pytorchbot retest this please |
zdevito
pushed a commit
to zdevito/ATen
that referenced
this pull request
Jul 24, 2018
Summary: Fixes: #9754 Maybe this could also make its way into 0.4.1, it is a severe debugging headache if you hit this... Pull Request resolved: pytorch/pytorch#9755 Reviewed By: ezyang Differential Revision: D8967178 Pulled By: zou3519 fbshipit-source-id: 151ed24e3a15a0c67014e411ac808fb893929a42
gchanan
added a commit
to gchanan/pytorch
that referenced
this pull request
Jul 25, 2018
pytorch#9755 broke this, but it was only tested if size zero dims were turned on.
gchanan
added a commit
to gchanan/pytorch
that referenced
this pull request
Jul 25, 2018
pytorch#9755 broke this, but it was only tested if size zero dims were turned on.
Contributor
|
Let's try not to reduce functionality (or question why and decide it's worth it); logsumexp worked on empty tensors before this and now it doesn't (see #9825). |
Collaborator
Author
|
@gchanan sorry! I was still dizzy from the non-inplacedness... |
facebook-github-bot
pushed a commit
that referenced
this pull request
Jul 25, 2018
Summary: #9755 broke this, but it was only tested if size zero dims were turned on (it can still happen even if that isn't turned on, because we support size [0] tensors). Pull Request resolved: #9825 Differential Revision: D8997303 Pulled By: gchanan fbshipit-source-id: 911dce112f73fad0f3980a7f4f9423df0f2d923d
zdevito
pushed a commit
to zdevito/ATen
that referenced
this pull request
Jul 25, 2018
Summary: pytorch/pytorch#9755 broke this, but it was only tested if size zero dims were turned on (it can still happen even if that isn't turned on, because we support size [0] tensors). Pull Request resolved: pytorch/pytorch#9825 Differential Revision: D8997303 Pulled By: gchanan fbshipit-source-id: 911dce112f73fad0f3980a7f4f9423df0f2d923d
jramseyer
pushed a commit
to jramseyer/pytorch
that referenced
this pull request
Jul 30, 2018
Summary: Fixes: pytorch#9754 Maybe this could also make its way into 0.4.1, it is a severe debugging headache if you hit this... Pull Request resolved: pytorch#9755 Reviewed By: ezyang Differential Revision: D8967178 Pulled By: zou3519 fbshipit-source-id: 151ed24e3a15a0c67014e411ac808fb893929a42
jramseyer
pushed a commit
to jramseyer/pytorch
that referenced
this pull request
Jul 30, 2018
Summary: pytorch#9755 broke this, but it was only tested if size zero dims were turned on (it can still happen even if that isn't turned on, because we support size [0] tensors). Pull Request resolved: pytorch#9825 Differential Revision: D8997303 Pulled By: gchanan fbshipit-source-id: 911dce112f73fad0f3980a7f4f9423df0f2d923d
goodlux
pushed a commit
to goodlux/pytorch
that referenced
this pull request
Aug 15, 2018
Summary: Fixes: pytorch#9754 Maybe this could also make its way into 0.4.1, it is a severe debugging headache if you hit this... Pull Request resolved: pytorch#9755 Reviewed By: ezyang Differential Revision: D8967178 Pulled By: zou3519 fbshipit-source-id: 151ed24e3a15a0c67014e411ac808fb893929a42
goodlux
pushed a commit
to goodlux/pytorch
that referenced
this pull request
Aug 15, 2018
Summary: pytorch#9755 broke this, but it was only tested if size zero dims were turned on (it can still happen even if that isn't turned on, because we support size [0] tensors). Pull Request resolved: pytorch#9825 Differential Revision: D8997303 Pulled By: gchanan fbshipit-source-id: 911dce112f73fad0f3980a7f4f9423df0f2d923d
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 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.
Fixes: #9754
Maybe this could also make its way into 0.4.1, it is a severe debugging headache if you hit this...