Add head_mask and decoder_head_mask to PyTorch LED (#9856)

* Add {decoder_,}head_mask to LED

* Fix create_custom_forward signatue in encoder

* Add head_mask to longformer

* Add head_mask to longformer to fix dependencies
of LED on Longformer.

* Not working yet

* Add mising one input in longofrmer_modeling.py

* make fix-copies

tests/test_modeling_longformer.py

@@ -273,7 +273,6 @@ class LongformerModelTester:
 @require_torch
 class LongformerModelTest(ModelTesterMixin, unittest.TestCase):
     test_pruning = False  # pruning is not supported
-    test_headmasking = False  # head masking is not supported
     test_torchscript = False
 
     all_model_classes = (