[optim] Widen the cases for defaulting to foreach #95820
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
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/95820
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit d9f191b: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
janeyx99
added
ciflow/trunk
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. [ghstack-poisoned]
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. [ghstack-poisoned]
janeyx99
requested review from
H-Huang,
awgu,
fegin,
kwen2501,
mrshenli,
rohan-varma,
wanchaol and
zhaojuanmao
as code owners
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. [ghstack-poisoned]
albanD
approved these changes
Mar 2, 2023
Collaborator
albanD
left a comment
albanD
left a comment
There was a problem hiding this comment.
Sounds good!
Did the doc PR landed? Does the phrasing on when we use foreach by default need to be changed?
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 |
Contributor
Author
Doc PR landed last week--phrasing could be minorly updated from "tensors" to "parameters" but not really significant. |
janeyx99
added a commit
to janeyx99/pytorch
that referenced
this pull request
Mar 2, 2023
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. Pull Request resolved: pytorch#95820 Approved by: https://github.com/albanD
This was referenced Mar 2, 2023
atalman
pushed a commit
that referenced
this pull request
Mar 2, 2023
* [optim] include nn.Parameter as foreach supported (#95811) This PR is a result of a realization that models are NOT subscribed to the foreach defaulting as have been claimed on our documentation for months now. BIG OOPS. Pull Request resolved: #95811 Approved by: https://github.com/albanD * [optim] Widen the cases for defaulting to foreach (#95820) Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. Pull Request resolved: #95820 Approved by: https://github.com/albanD
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 5, 2023
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. Pull Request resolved: pytorch/pytorch#95820 Approved by: https://github.com/albanD
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 5, 2023
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. Pull Request resolved: pytorch/pytorch#95820 Approved by: https://github.com/albanD
pruthvistony
pushed a commit
to ROCm/pytorch
that referenced
this pull request
May 3, 2023
* [optim] include nn.Parameter as foreach supported (pytorch#95811) This PR is a result of a realization that models are NOT subscribed to the foreach defaulting as have been claimed on our documentation for months now. BIG OOPS. Pull Request resolved: pytorch#95811 Approved by: https://github.com/albanD * [optim] Widen the cases for defaulting to foreach (pytorch#95820) Big OOP correction continued. Also added a test this time to verify the defaulting was as expected. The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that _all_ tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach _can_ handle cpu tensors, this should not introduce breakage. Pull Request resolved: pytorch#95820 Approved by: https://github.com/albanD
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.
Big OOP correction continued. Also added a test this time to verify the defaulting was as expected.
The key here is realizing that the grouping for foreach already assumes that the non-param tensorlists follow suit in dtype and device, so it is too narrow to check that all tensors were on CUDA. The main leeway this allowed was state_steps, which are sometimes cpu tensors. Since foreach can handle cpu tensors, this should not introduce breakage.
Stack from ghstack (oldest at bottom):