Hi @hyperfraise!

Thank you for your pull request and welcome to our community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file.

In order for us to review and merge your code, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

If you have received this in error or have any questions, please contact us at cla@fb.com. Thanks!

💊 CI failures summary and remediations

As of commit c11dd20 (more details on the Dr. CI page):

• 1/2 failures possibly* introduced in this PR
  • 1/1 non-scanned failure(s)
• 1/2 broken upstream at merge base d90d724 from Mar 02 until Mar 03

🚧 1 fixed upstream failure:

These were probably caused by upstream breakages that were already fixed.

git fetch https://github.com/pytorch/pytorch viable/strict
git rebase FETCH_HEAD

• pytorch_windows_vs2019_py36_cuda10.1_test1 from Mar 02 until Mar 03 (816646b - 59b2b8b)
  • 🔁 rerun

ci.pytorch.org: 1 failed

• Failed: pr/pytorch-linux-bionic-rocm4.0.1-py3.6

Codecov Report

Merging #50003 (68579f6) into master (0c45533) will decrease coverage by 0.21%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master   #50003      +/-   ##
==========================================
- Coverage   80.80%   80.59%   -0.22%     
==========================================
  Files        1972     1957      -15     
  Lines      216286   214523    -1763     
==========================================
- Hits       174774   172888    -1886     
- Misses      41512    41635     +123     

Could you separate the formatting changes and the feature?

can you update the test here: https://github.com/pytorch/pytorch/blob/master/test/quantization/test_quantize_fx.py#L397 (add a 3 to the list)

Could you separate the formatting changes and the feature?

Thanks for the review ! They are separated in different commits; Do you mean opening a new PR ?

can you update the test here: https://github.com/pytorch/pytorch/blob/master/test/quantization/test_quantize_fx.py#L397 (add a 3 to the list)

I will.

Could you separate the formatting changes and the feature?

Thanks for the review ! They are separated in different commits; Do you mean opening a new PR ?

I see. ideally it should be a new PR I think, but separate commits probably works as well.

Looks good, thanks for adding this!

Wait there's the rocm test I can't pass, do you see why ? I can't figure why

@albanD @mingzhe09088 @mrshenli @pritamdamania87 @rohan-varma @zhaojuanmao @jerryzh168 @ezyang @fmassa @glaringlee @soumith
It'd be great if you could review this, please, I'm having to solve conflicts quite often.

cc @jerryzh168 Is this PR ready to land?

Sorry @hyperfraise for the late response, there were still some internal errors (in phabricator) in aten/src/ATen/native/metal/MetalPrepackOpRegister.cpp:

unknown type name 'Conv3dOpContext'; did you mean 'Conv2dOpContext'?
fbcode/caffe2/aten/src/ATen/native/metal/MetalPrepackOpContext.h:20:7: 'Conv2dOpContext' declared here
class Conv2dOpContext : public torch::jit::CustomClassHolder {


Linter (CLANGTIDY)raised a clang-diagnostic-error error online 120-
no member named 'prod_intlist' in namespace 'at'

cc @xta0 is there a way that we can compile metal in OSS to make sure it works?

Hi @hyperfraise could you revert the metal changes? we can add it internal if it's needed. the ci/clang errors are only visible internally so it's hard to figure out what to do in oss.

Hi @hyperfraise could you revert the metal changes? we can add it internal if it's needed. the ci/clang errors are only visible internally so it's hard to figure out what to do in oss.

Ofc ! Let's drop them.

I rebased with your suggestions. It seems there's an unrelated failure in the CI, so you can still review.

@jerryzh168 merged this pull request in f918597.

@hyperfraise sorry for the long delay, it's finally landed!

@hyperfraise sorry for the long delay, it's finally landed!

Nothing undeserved ^^ Glad to open the next PR if the issue mandates one.

Thanks for your help.