enable test_torch_compile_recompilation_and_graph_break on XPU

Signed-off-by: YAO Matrix <[email protected]>

Author: yao-matrix

tests/models/test_modeling_common.py

@@ -1759,7 +1759,7 @@ def tearDown(self):
         gc.collect()
         backend_empty_cache(torch_device)
 
-    @require_torch_gpu
+    @require_torch_accelerator
     @require_torch_2
     @is_torch_compile
     @slow