Fix #92108 (#92870)
You can easily test this by adding
```
@patch.object(config.triton, "convolution", "triton")
```
to test_convolution1 but it takes a long time to autotune so
I don't want to add it to the unit tests.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/92870
Approved by: https://github.com/albanD
