mikaylagawarecki

@pytorchbot merge

@pytorchbot merge

@pytorchbot merge

@pytorchbot merge

Error seems to come from [here](https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cudnn/Conv_v8.cpp#L650) not sure why it is resolved by manual installation of torchtriton

Hi @nimashiri I was unable to reproduce on 1.13.1, could you upgrade your torch version and retry please? In general, I see that you have filed many issues regarding SegFaults...

Hi @nimashiri, I have repurposed this issue as a tracking issue for the segfault/floating point exception issues that you filed yesterday, with the drive link from each issue linked above....

The error on the latest master is `"atan2_cpu" not implemented for 'ComplexFloat'`, would you like to request complexFloat support for atan2?

The CI failure looks unrelated to me, I just reran it. Could you add some tests to make sure the rounding behavior across gridsampling ops for 0.5 is consistent please

Of course, there are existing tests for grid_sample in [test/test_nn.py](https://github.com/pytorch/pytorch/blob/master/test/test_nn.py), you can add a test there Let me know if you have any questions! Edit: @leimao before you do this,...