conv-bn folding should be used for CNN-based Vision Models when AMP is used with oneDNN Graph #2535
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
…sed with AMP By default, conv-bn folding isn't done with CNN based models when AMP is used with oneDNN Graph. `torch.fx.experimental.optimize.fuse` should be used for such models
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/2535
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 812b863 with merge base 3a769b0 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Member
|
This seems fine, I'm a bit confused by |
msaroufim
approved these changes
Aug 25, 2023
Contributor
Author
|
Thanks for reviewing, @msaroufim! :) |
svekars
reviewed
Aug 25, 2023
Co-authored-by: Svetlana Karslioglu <svekars@meta.com>
svekars
approved these changes
Aug 25, 2023
mikaylagawarecki
pushed a commit
to mikaylagawarecki/tutorials
that referenced
this pull request
Sep 18, 2023
…s used with oneDNN Graph (pytorch#2535) * Update tuning guide to reflect folding conv-bn when oneDNN Graph is used with AMP By default, conv-bn folding isn't done with CNN based models when AMP is used with oneDNN Graph. `torch.fx.experimental.optimize.fuse` should be used for such models --------- Co-authored-by: Svetlana Karslioglu <svekars@meta.com>
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Outdated suggestions cannot be applied.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.
By default, conv-bn folding isn't done with CNN based models when AMP is used with oneDNN Graph.
torch.fx.experimental.optimize.fuseshould be used for such models