[autograd][docs] Add more details on why save_for_backward is important in extending autograd note #153005
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
…nt in extending autograd note [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/153005
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 1 Unrelated FailureAs of commit 73394bc with merge base e2c7ae5 ( NEW FAILURE - The following job has failed:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
… is important in extending autograd note" [ghstack-poisoned]
stas00
reviewed
May 7, 2025
docs/source/notes/extending.rst
Outdated
Comment on lines
115
to
124
| Saving tensors via ``save_for_backward`` allows the autograd engine to clear | ||
| them as soon as the backward computation of the ``autograd.Function`` completes. | ||
| (If a tensor is stored directly on ``ctx`` | ||
| it will unnecessarily remain alive for the lifetime of the autograd graph -- | ||
| typically until the end of the iteration.) Using ``save_for_backward`` | ||
| also helps avoid certain reference cycles, (e.g., since the tensor | ||
| output of the ``autograd.Function`` itself keeps a reference to the ctx). | ||
| Saving via ``save_for_backward`` is also important for compatibility with | ||
| features like activation checkpointing and offloading that rely on | ||
| :class:`torch.autograd.graph.saved_tensors_hooks`. |
Contributor
There was a problem hiding this comment.
A possible readability improvement - not sure if I formatted it correctly - been a while since I touched rst.
Suggested change
| Saving tensors via ``save_for_backward`` allows the autograd engine to clear | |
| them as soon as the backward computation of the ``autograd.Function`` completes. | |
| (If a tensor is stored directly on ``ctx`` | |
| it will unnecessarily remain alive for the lifetime of the autograd graph -- | |
| typically until the end of the iteration.) Using ``save_for_backward`` | |
| also helps avoid certain reference cycles, (e.g., since the tensor | |
| output of the ``autograd.Function`` itself keeps a reference to the ctx). | |
| Saving via ``save_for_backward`` is also important for compatibility with | |
| features like activation checkpointing and offloading that rely on | |
| :class:`torch.autograd.graph.saved_tensors_hooks`. | |
| Saving tensors via ``save_for_backward``: | |
| 1. allows the autograd engine to clear | |
| them as soon as the backward computation of the ``autograd.Function`` completes. | |
| (If a tensor is stored directly on ``ctx`` | |
| it will unnecessarily remain alive for the lifetime of the autograd graph -- | |
| typically until the end of the iteration.) | |
| 2. it helps avoid certain reference cycles, (e.g., since the tensor | |
| output of the ``autograd.Function`` itself keeps a reference to the ctx). | |
| 3. it is also important for compatibility with | |
| features like activation checkpointing and offloading that rely on | |
| :class:`torch.autograd.graph.saved_tensors_hooks`. |
Contributor
Author
There was a problem hiding this comment.
Thanks, updated. Will double check the rendered results
… is important in extending autograd note" cc stas00 [ghstack-poisoned]
Contributor
Author
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Collaborator
Merge failedReason: 1 jobs have failed, first few of them are: trunk / win-vs2022-cpu-py3 / test (default, 1, 3, ephemeral.windows.4xlarge.nonephemeral) Details for Dev Infra teamRaised by workflow job |
Contributor
Author
|
@pytorchbot merge -i |
Collaborator
Merge startedYour change will be merged while ignoring the following 2 checks: trunk / macos-py3-arm64 / test (default, 3, 3, macos-m1-stable), trunk / win-vs2022-cpu-py3 / test (default, 1, 3, ephemeral.windows.4xlarge.nonephemeral) Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Collaborator
|
Starting merge as part of PR stack under #153094 |
pytorchmergebot
pushed a commit
that referenced
this pull request
May 12, 2025
Fixes #152773 Pull Request resolved: #153094 Approved by: https://github.com/albanD ghstack dependencies: #153005
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.
Stack from ghstack (oldest at bottom):
cc @stas00