A recipe for showcasing torch.export flow for 4 models, with unique challenges and solutions #3180
Conversation
This file contains 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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/3180
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 71c468f with merge base f7d06b6 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
svekars
reviewed
Dec 10, 2024
svekars
reviewed
Dec 10, 2024
Co-authored-by: Svetlana Karslioglu <svekars@meta.com>
Co-authored-by: Svetlana Karslioglu <svekars@meta.com>
ydwu4
reviewed
Dec 11, 2024
|
|
||
| TORCH_LOGS="graph_breaks" python <file_name>.py | ||
|
|
||
| You will need to modify your program to get rid of graph breaks. Once resolved, you are ready to export the model. |
There was a problem hiding this comment.
There is an alternative of fixing all the graph breaks, which is export with strict=False and strict=False will become the default setting soon cc @avikchaudhuri. So it may worth mentioning this up front (I see the non-strict example below but just feel it worth emphasizing)
There was a problem hiding this comment.
discussed offline that will be addressing this in a different tutorial in H1
angelayi
reviewed
Jan 15, 2025
Co-authored-by: Angela Yi <angelayi@meta.com>
angelayi
reviewed
Jan 17, 2025
angelayi
approved these changes
Jan 17, 2025
Co-authored-by: Angela Yi <angelayi@meta.com>
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Outdated suggestions cannot be applied.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.
Fixes #3174
Description
The gaming community cares about certain classes of models like pose estimation, instance segmentation, video classification. When we try to export OSS implementations of these models, we run into unique challenges with torch.export
Currently, we have tutorials showing usage of export and talking about the core export-related concepts to keep in mind with simple examples. We also have ExportDB which has information on unsupported constructs with simple examples. However, practically, when running export on many models, its not very clear how does once go about addressing the issues.
This tutorial aims to do the reverse. Pick 4 models which are popular, try to export them, show the errors we run into and how do we solve them. The problems being solved are generic enough to be applicable to a range of models.
Checklist
cc @avikchaudhuri @gmagogsfm @zhxchen17 @tugsbayasgalan @angelayi @suo @ydwu4