Closed
Conversation
#39611 unified signatures of some ops taking TensorOptions arguments by making them optional. That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR). However, it also changed the default from `pin_memory=False` to `pin_memory=None`, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch, because we materialize default values when exporting. This is bad. This PR reverts that particular part of #39611 to revert the FC breakage. Differential Revision: [D22461661](https://our.internmc.facebook.com/intern/diff/D22461661/) [ghstack-poisoned]
smessmer
added a commit
that referenced
this pull request
Jul 9, 2020
#39611 unified signatures of some ops taking TensorOptions arguments by making them optional. That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR). However, it also changed the default from `pin_memory=False` to `pin_memory=None`, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch, because we materialize default values when exporting. This is bad. This PR reverts that particular part of #39611 to revert the FC breakage. Differential Revision: [D22461661](https://our.internmc.facebook.com/intern/diff/D22461661/) ghstack-source-id: 107456654 Pull Request resolved: #41198
bhosmer
approved these changes
Jul 9, 2020
#39611 unified signatures of some ops taking TensorOptions arguments by making them optional. That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR). However, it also changed the default from `pin_memory=False` to `pin_memory=None`, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch, because we materialize default values when exporting. This is bad. This PR reverts that particular part of #39611 to revert the FC breakage. Differential Revision: [D22461661](https://our.internmc.facebook.com/intern/diff/D22461661/) [ghstack-poisoned]
#39611 unified signatures of some ops taking TensorOptions arguments by making them optional. That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR). However, it also changed the default from `pin_memory=False` to `pin_memory=None`, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch, because we materialize default values when exporting. This is bad. This PR reverts that particular part of #39611 to revert the FC breakage. Differential Revision: [D22461661](https://our.internmc.facebook.com/intern/diff/D22461661/) [ghstack-poisoned]
smessmer
added a commit
that referenced
this pull request
Jul 9, 2020
Pull Request resolved: #41198 #39611 unified signatures of some ops taking TensorOptions arguments by making them optional. That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR). However, it also changed the default from `pin_memory=False` to `pin_memory=None`, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch, because we materialize default values when exporting. This is bad. This PR reverts that particular part of #39611 to revert the FC breakage. ghstack-source-id: 107475024 Differential Revision: [D22461661](https://our.internmc.facebook.com/intern/diff/D22461661/)
Contributor
|
This pull request has been merged in 5f6c6ed. |
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
4 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:
#39611 unified signatures of some ops taking TensorOptions arguments by making them optional.
That has FC implications but only for models writting with a PyTorch version after that version (see explanation in description of that PR).
However, it also changed the default from
pin_memory=Falsetopin_memory=None, which actually breaks FC for preexisting models too if they're re-exported with a newer PyTorch,because we materialize default values when exporting. This is bad.
This PR reverts that particular part of #39611 to revert the FC breakage.
Differential Revision: D22461661