[Tests] Add Common Test for Training + Fix a couple of bugs (#8415)

* add training tests

* correct longformer

* fix docs

* fix some tests

* fix some more train tests

* remove ipdb

* fix multiple edge case model training

* fix funnel and prophetnet

* clean gpt models

* undo renaming of albert

tests/test_modeling_electra.py

@@ -24,7 +24,10 @@ from .test_modeling_common import ModelTesterMixin, ids_tensor, random_attention
 
 
 if is_torch_available():
+    import torch
+
     from transformers import (
+        MODEL_FOR_PRETRAINING_MAPPING,
         ElectraConfig,
         ElectraForMaskedLM,
         ElectraForMultipleChoice,
@@ -285,6 +288,17 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase):
         else ()
     )
 
+    # special case for ForPreTraining model
+    def _prepare_for_class(self, inputs_dict, model_class, return_labels=False):
+        inputs_dict = super()._prepare_for_class(inputs_dict, model_class, return_labels=return_labels)
+
+        if return_labels:
+            if model_class in MODEL_FOR_PRETRAINING_MAPPING.values():
+                inputs_dict["labels"] = torch.zeros(
+                    (self.model_tester.batch_size, self.model_tester.seq_length), dtype=torch.long, device=torch_device
+                )
+        return inputs_dict
+
     def setUp(self):
         self.model_tester = ElectraModelTester(self)
         self.config_tester = ConfigTester(self, config_class=ElectraConfig, hidden_size=37)