StopIteration Error in torch.fx tutorial with TransformerEncoderLayer

[lkp411]

I’m trying to run the fx profiling tutorial in tutorials/fx_profiling_tutorial.py at master · pytorch/tutorials · GitHub 1 on a single nn.TransformerEncoderLayer as opposed to the resnet in the example and I keep running into a StopIteration error. Why is this happening? All I did was replace the resnet with a transformer encoder layer.

cc @eellison @suo @gmagogsfm @jamesr66a @msaroufim @SherlockNoMad @sekyondaMeta @svekars @carljparker @NicolasHug @kit1980 @subramen

[LujiAnna]

/assigntome

[LujiAnna]

Hi @lkp411 ,
Do you mind sharing your code, when you were using nn.TransformerEncoderLayer instead of ResNet, to see which part produces the StopIteration error ?

[LujiAnna]

/assigntome

[lkp411]

Hey there, I unfortunately don't have this code anymore as it was over 18 months ago.

…

On Tue, Jun 13, 2023 at 1:17 AM anna ***@***.***> wrote: /assigntome — Reply to this email directly, view it on GitHub <https://urldefense.com/v3/__https://github.com/pytorch/tutorials/issues/1705*issuecomment-1588784862__;Iw!!LIr3w8kk_Xxm!roN8974SpdqKCox67D8TaW_DUZfdMBNev5nyG4rm-4oxbYCTSVCDrmsWRx3b53wXsHS4Zpf27p-JVa27OiPx9qUe$>, or unsubscribe <https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/ALW4JPZLTDGEUOCVGTK4F3TXLAOZNANCNFSM5FU6CC6A__;!!LIr3w8kk_Xxm!roN8974SpdqKCox67D8TaW_DUZfdMBNev5nyG4rm-4oxbYCTSVCDrmsWRx3b53wXsHS4Zpf27p-JVa27OvT9IrGD$> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[github-actions]

The issue is already assigned. Please pick an opened and unnasigned issue with the docathon-h1-2023 label.

[svekars]

This issue has been unassigned due to inactivity. If you are still planning to work on this, you can still send a PR referencing this issue.