This repository was archived by the owner on Aug 1, 2025. It is now read-only.
Add lowering for adaptive avg pool #962
Merged
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
|
Hm, I would think they reduce to 1x1 output and thus could be replaced by sum, not no-op? |
ngimel
reviewed
Aug 23, 2022
torchinductor/lowering.py
Outdated
| # no-op if the same input and output | ||
| if h_in == h_out and w_in == w_out: | ||
| # TODO: do I need to copy ? _to_copy does not | ||
| return x |
There was a problem hiding this comment.
better do a copy, yes (to make sure that previously run functionalization is still valid)
There was a problem hiding this comment.
torchdynamo/torchinductor/lowering.py
Line 915 in 4b212e1
The output size is given, not the kernel size which differs from the other apis... When the output_size == input_size for adaptive average pooling, that means the kernel size will be [1, 1] and you're doing a 1-element average (no-op). |
jansel
reviewed
Aug 23, 2022
torchinductor/lowering.py
Outdated
| # no-op if the same input and output | ||
| if h_in == h_out and w_in == w_out: | ||
| # TODO: do I need to copy ? _to_copy does not | ||
| return x |
There was a problem hiding this comment.
torchdynamo/torchinductor/lowering.py
Line 915 in 4b212e1
jansel
approved these changes
Aug 23, 2022
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 25, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 25, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 25, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 25, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). ghstack-source-id: 3875878 Pull Request resolved: #84062
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). ghstack-source-id: 0473704 Pull Request resolved: #84062
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 26, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 27, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 27, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 27, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). ghstack-source-id: 09e090f Pull Request resolved: #84062
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 28, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here (I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright). ghstack-source-id: 656a53e Pull Request resolved: #84062
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 28, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
lezcano
added a commit
to pytorch/pytorch
that referenced
this pull request
Aug 28, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. [ghstack-poisoned]
pytorchmergebot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Aug 28, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. Pull Request resolved: #84062 Approved by: https://github.com/jansel
pytorchmergebot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Aug 29, 2022
This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. Pull Request resolved: #84062 Approved by: https://github.com/jansel
facebook-github-bot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Aug 29, 2022
Summary: This was already implemented as a lowering in pytorch/torchdynamo#962. I'm putting the idea up here ~(I haven't even run this code, so it surely has *many* issues, but I reckon the general idea should hopefully be alright).~ The tests now pass and I corrected the issues that the first implementation had. Pull Request resolved: #84062 Approved by: https://github.com/jansel Test Plan: contbuild & OSS CI, see https://hud.pytorch.org/commit/pytorch/pytorch/9fc02f6bc558f26a460241cbeaea915ea2b41005 Reviewed By: weiwangmeta Differential Revision: D39091747 fbshipit-source-id: f984c729e355d2d1a4abe0c01a97b30fbdca95b7
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
The two occurrences in torchbench of this op are both no-ops lol. When you disable the no-op optimization, the op is ~6x faster (again, for those two sets of inputs).