Add tie_weights() to LM heads and set bias in set_output_embeddings() (#28948)

* Add tie_weights() to LM heads and set bias in set_output_embeddings()

The bias were not tied correctly in some LM heads, and this change should fix that.

* Moving test_save_and_load_low_cpu_mem_usage to ModelTesterMixin

* Adding _tie_weights() to MPNet and Vilt

* Skip test for low cpu mem usage for Deta/DeformableDetr since they cannot init on meta device

* Rename to test name to save_load to match the convention

tests/models/deformable_detr/test_modeling_deformable_detr.py

@@ -564,6 +564,10 @@ class DeformableDetrModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineT
                         msg=f"Parameter {name} of model {model_class} seems not properly initialized",
                     )
 
+    @unittest.skip("Cannot be initialized on meta device as some weights are modified during the initialization")
+    def test_save_load_low_cpu_mem_usage(self):
+        pass
+
     def test_two_stage_training(self):
         model_class = DeformableDetrForObjectDetection
         config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()

tests/models/deta/test_modeling_deta.py

@@ -520,6 +520,10 @@ class DetaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
                         msg=f"Parameter {name} of model {model_class} seems not properly initialized",
                     )
 
+    @unittest.skip("Cannot be initialized on meta device as some weights are modified during the initialization")
+    def test_save_load_low_cpu_mem_usage(self):
+        pass
+
 
 TOLERANCE = 1e-4