It is more about using nightly pytorch where the CUDA version of trition it seems harcorded:
I am on the same commit as our CI:
How we are going to still release working (torch.compile
) cuda 11.8 nightly wheels?
It is more about using nightly pytorch where the CUDA version of trition it seems harcorded:
I am on the same commit as our CI:
How we are going to still release working (torch.compile
) cuda 11.8 nightly wheels?