Add XPU type for work-around -inf mask causing sdpa NaN issue in modeling files (#35647)

* add xpu for unmask

* change modular for generated matching

* add lastest modeling for helium

examples/modular-transformers/modeling_dummy.py

@@ -639,7 +639,7 @@ class DummyModel(DummyPreTrainedModel):
         if (
             self.config._attn_implementation == "sdpa"
             and attention_mask is not None
-            and attention_mask.device.type == "cuda"
+            and attention_mask.device.type in ["cuda", "xpu"]
             and not output_attentions
         ):
             # Attend to all tokens in fully masked rows in the causal_mask, for example the relevant first rows when