Symbolic shape tests don't actually test that arguments aren't specialized #120911
Labels
module: dynamic shapes
module: tests
Issues related to tests (not the torch.testing module)
oncall: pt2
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
馃悰 Describe the bug
For example, this means that #120776 didn't cause an xfail to be removed. This is suboptimal.
The difficulty here is that we need to actually annotate which inputs are expected to be specialized and which are not.
Prior to us switching to automatically specializing, we could sort of get some coverage info because if someone called
.size()
this would hard error, but now it doesn't anymore.Versions
main
cc @mruberry @ZainRizvi @msaroufim @bdhirsh @anijain2305 @zou3519 @chauhang
The text was updated successfully, but these errors were encountered: