Apply suggestions from code review

Co-authored-by: Ella Charlaix <80481427+echarlaix@users.noreply.github.com>

optimum/intel/utils/modeling_utils.py

@@ -184,7 +184,7 @@ def recursive_to_device(value, device):
     return value
 
 
-def setattr_from_module(new_module, module):
+def _setattr_from_module(new_module, module):
     for k, v in module.__dict__.items():
         setattr(new_module, k, v)
     for k, v in module.__class__.__dict__.items():

tests/ipex/test_modeling.py

@@ -242,10 +242,6 @@ def test_pipeline(self, model_arch):
 
     # High optimized model llama is not supported assisted decoding for now.
     @parameterized.expand(SUPPORTED_ARCHITECTURES)
-    @unittest.skipIf(
-        is_ipex_version(">=", "2.3.0"),
-        reason="CPU IPEXModel does not support assisted decoding when ipex version >= 2.3.0",
-    )
     def test_assisted_decoding(self, model_arch):
         if model_arch == "llama2":
             return
@@ -301,10 +297,6 @@ def test_ipex_patching_beam_search(self, test_name, model_arch, use_cache):
                 self.assertIsInstance(outputs, torch.Tensor)
                 self.assertTrue(torch.equal(outputs, transformers_outputs))
 
-    @unittest.skipIf(
-        is_ipex_version(">=", "2.3.0"),
-        reason="CPU IPEXModel only supports with past_key_values for ipex version >= 2.3.0",
-    )
     def test_compare_with_and_without_past_key_values(self):
         model_id = "echarlaix/tiny-random-gpt2-torchscript"
         tokenizer = AutoTokenizer.from_pretrained(model_id)