Fix cuda/cpu check on NoneType (#88854) (#90068)

Summary: Fix cuda/cpu check on NoneType

Test Plan: sabdcastle/ github CI/CD

Differential Revision: D41203955

Pull Request resolved: #88854
Approved by: https://github.com/drisspg, https://github.com/ngimel

Co-authored-by: Michael Gschwind <[email protected]>

Author: weiwangmeta

Diff for: torch/nn/modules/activation.py

@@ -1116,7 +1116,7 @@ def forward(self, query: Tensor, key: Tensor, value: Tensor, key_padding_mask: O
             # generator expressions.
             if torch.overrides.has_torch_function(tensor_args):
                 why_not_fast_path = "some Tensor argument has_torch_function"
-            elif not all([(x.is_cuda or 'cpu' in str(x.device)) for x in tensor_args]):
+            elif not all([(x is None or x.is_cuda or 'cpu' in str(x.device)) for x in tensor_args]):
                 why_not_fast_path = "some Tensor argument is neither CUDA nor CPU"
             elif torch.is_grad_enabled() and any([x is not None and x.requires_grad for x in tensor_args]):
                 why_not_fast_path = ("grad is enabled and at least one of query or the "