Skip to content

Conversation

@pmeier
Copy link
Contributor

@pmeier pmeier commented Oct 25, 2022

I often encounter a warning like

test/test_prototype_transforms_functional.py::TestKernels::test_scripted_vs_eager[cpu-five_crop_image_tensor-07]
  /home/philip/git/pytorch/torchvision/test/test_prototype_transforms_functional.py:98: UserWarning: operator() profile_node %373 : int[] = prim::profile_ivalue(%371)
   does not have profile information (Triggered internally at ../torch/csrc/jit/codegen/cuda/graph_fuser.cpp:105.)
    actual = kernel_scripted(*args, **kwargs)

when running the prototype transforms tests locally. The numbers, i.e. %373 and %371, from above are changing. So far I haven't figured out a pattern or the like.

In #6785, we added the first ignore for it and #6783 added another. I'm not keen on fixing this warning over and over if it adds no information. Thus, this PR ignores it in all of the JIT tests on the prototype transforms side.

cc @vfdev-5 @datumbox @bjuncek

Copy link
Contributor

@datumbox datumbox left a comment

Choose a reason for hiding this comment

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

LGTM, thanks!

@pmeier pmeier merged commit 8e0e715 into pytorch:main Oct 26, 2022
@pmeier pmeier deleted the fix-jit-warning branch October 26, 2022 06:43
facebook-github-bot pushed a commit that referenced this pull request Oct 27, 2022
Summary:
* ignore git warning 'globally'

* improve comment

Reviewed By: YosuaMichael

Differential Revision: D40722905

fbshipit-source-id: 7c0d94689cda92d8193c0914f7fa423cbe4505fb
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants