I believe that PyTorch already uses Triton; I recently tried to do torch.compile on a Windows machine and it did not work because the inductor backend relies on

I believe that PyTorch already uses Triton; I recently tried to do torch.compile on a Windows machine and it did not work because the inductor backend relies on Triton.