[tests] use torch_device instead of auto for model testing (huggi…

…ngface#29531)

* use torch_device

* skip for XPU

* Update tests/generation/test_utils.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

---------

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

tests/generation/test_utils.py

@@ -1073,6 +1073,9 @@ def test_beam_search_generate_dict_outputs_use_cache(self):
     @require_torch_multi_accelerator
     def test_model_parallel_beam_search(self):
         for model_class in self.all_generative_model_classes:
+            if "xpu" in torch_device:
+                return unittest.skip("device_map='auto' does not work with XPU devices")
+
             if model_class._no_split_modules is None:
                 continue