8000 pin update by lucylq · Pull Request #7273 · pytorch/executorch · GitHub
[go: up one dir, main page]

Skip to content

pin update #7273

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Dec 20, 2024
Merged

pin update #7273

merged 2 commits into from
Dec 20, 2024

Conversation

lucylq
Copy link
Contributor
@lucylq lucylq commented Dec 10, 2024

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Dec 10, 2024
Copy link
pytorch-bot bot commented Dec 10, 2024

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/7273

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit 64635c1 with merge base 18142f7 (image):
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@lucylq lucylq marked this pull request as ready for review December 13, 2024 17:02
@mergennachin mergennachin self-requested a review December 13, 2024 17:15
@lucylq
Copy link
Contributor Author
lucylq commented Dec 13, 2024

Fix related to training IR migration for arm unittest:
#7275

@lucylq lucylq force-pushed the ay.pin-update branch 3 times, most recently from 7798e48 to e14318f Compare December 16, 2024 22:55
@lucylq
Copy link
Contributor Author
lucylq commented Dec 16, 2024

QNN failure potentially due to functionalization changes:

# TODO: this is temporary and export_for_training doesn't work with qnn either. We need a
# functional graph. See issue https://github.com/pytorch/executorch/pull/4627 for more details
exported_module = torch.export.export(
self.model,
self.example_inputs,
self.example_kwarg_inputs,
dynamic_shapes=dynamic_shape,
strict=True,

@@ -193,6 +193,7 @@ def export(self) -> "LLMEdgeManager":
dynamic_shapes=dynamic_shape,
strict=True,
)
exported_module.run_decompositions({})
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You need to re-assign exported_module, it is not an inplace update function.

@lucylq lucylq force-pushed the ay.pin-update branch 2 times, most recently from 7ae005e to 844fa60 Compare December 17, 2024 19:38
@lucylq
Copy link
Contributor Author
lucylq commented Dec 17, 2024

fix for arm test: pytorch/pytorch#143376

@lucylq lucylq force-pushed the ay.pin-update branch 6 times, most recently from f3845fe to a3d70a5 Compare December 18, 2024 20:02
@swolchok
Copy link
Contributor
swolchok commented Dec 18, 2024

test-arm-backend-delegation failure appears to be a known problem fixed by @digantdesai in pytorch/pytorch#143376 14 hours ago; probably need to roll the version forward? (might also need to wait a day; not sure if it's hit a nightly yet or not)

@lucylq
Copy link
Contributor Author
lucylq commented Dec 18, 2024

test-arm-backend-delegation failure appears to be a known problem fixed by @digantdesai in pytorch/pytorch#143376 14 hours ago; probably need to roll the version forward? (might also need to wait a day; not sure if it's hit a nightly yet or not)

Yeah - looks like it's in the 2024-12-18 nightly, updating: https://github.com/pytorch/pytorch/blob/2ea4b56ec872424e486c4fe2d55da061067a2ed3/torch/fx/passes/infra/partitioner.py#L184

@lucylq lucylq force-pushed the ay.pin-update branch 2 times, most recently from f40a8b9 to d3b0371 Compare December 18, 2024 21:46
8000
@lucylq
Copy link
Contributor Author
lucylq commented Dec 18, 2024

adding patch for qnn to use old export flow: #7370

verbose=self.verbose,
)

override_export_behaviour = contextlib.nullcontext()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you leave a TODO here and mention the issue you filed?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ah - I left it at the top of the block.

@lucylq
Copy link
Contributor Author
lucylq commented Dec 20, 2024

pytorch/pytorch#143423

Pytorch nightly now sets _GLIBCXX_USE_CXX11_ABI by default. Think this is causing the errors in llava runner.

@lucylq lucylq force-pushed the ay.pin-update branch 2 times, most recently from b8256dc to c53b550 Compare December 20, 2024 00:51
@lucylq lucylq requested a review from huydhn December 20, 2024 00:52
@@ -298,7 +298,6 @@ prebuilt_cxx_library(
name = "libtorch",
shared_lib = ":libtorch_gen[libtorch]",
exported_preprocessor_flags = [
"-D_GLIBCXX_USE_CXX11_ABI=0", # `libtorch` is built without CXX11_ABI so any target depends on it need to use the same build config.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's see if this works. @atalman Do you know if we have switch libtorch build to CXX11_ABI yet? I don't see it in pytorch/pytorch#143423 yet

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All the tests are green! Good to land? @huydhn

@lucylq lucylq merged commit 59a2594 into main Dec 20, 2024
81 checks passed
@lucylq lucylq deleted the ay.pin-update branch January 24, 2025 19:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. topic: not user facing
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants
0