Modify resize_token_embeddings to ensure output type is same as input (#31979)

* Change resize_token_embeddings to make it return same Class that is passed to it

* Add explanatory comment as requested in review

* Add explanatory comments for add resizing function in lxmert

* Add comment for padding_idx and moving _resize_bias in lxmert to LxmertForPreTraining

---------

Co-authored-by: Prashanth Sateesh <prasatee@Prashanths-MBP.attlocal.net>
Co-authored-by: Prashanth Sateesh <prasatee@Prashanths-MacBook-Pro.local>

tests/test_modeling_common.py

@@ -1755,6 +1755,8 @@ class ModelTesterMixin:
             config = copy.deepcopy(original_config)
             model = model_class(config)
             model.to(torch_device)
+            model_embed_pre_resize = model.get_input_embeddings()
+            type_model_embed_pre_resize = type(model_embed_pre_resize)
 
             if self.model_tester.is_training is False:
                 model.eval()
@@ -1774,6 +1776,9 @@ class ModelTesterMixin:
             self.assertEqual(new_model_vocab_size, model_vocab_size + 10)
             # Check that it actually resizes the embeddings matrix
             self.assertEqual(model_embed.weight.shape[0], cloned_embeddings.shape[0] + 10)
+            # Check to make sure the type of embeddings returned post resizing is same as type of input
+            type_model_embed_post_resize = type(model_embed)
+            self.assertEqual(type_model_embed_pre_resize, type_model_embed_post_resize)
             # Check that the model can still do a forward pass successfully (every parameter should be resized)
             model(**self._prepare_for_class(inputs_dict, model_class))