Add DeiT (PyTorch) (#11056)

* First draft of deit

* More improvements

* Remove DeiTTokenizerFast from init

* Conversion script works

* Add DeiT to ViT conversion script

* Add tests, add head model, add support for deit in vit conversion script

* Update model checkpoint names

* Update image_mean and image_std, set resample to bicubic

* Improve docs

* Docs improvements

* Add DeiTForImageClassificationWithTeacher to init

* Address comments by @sgugger

* Improve feature extractors

* Make fix-copies

* Minor fixes

* Address comments by @patil-suraj

* All models uploaded

* Fix tests

* Remove labels argument from DeiTForImageClassificationWithTeacher

* Fix-copies, style and quality

* Fix tests

* Fix typo

* Multiple docs improvements

* More docs fixes

tests/test_configuration_common.py

@@ -20,14 +20,16 @@ import tempfile
 
 
 class ConfigTester(object):
-    def __init__(self, parent, config_class=None, **kwargs):
+    def __init__(self, parent, config_class=None, has_text_modality=True, **kwargs):
         self.parent = parent
         self.config_class = config_class
+        self.has_text_modality = has_text_modality
         self.inputs_dict = kwargs
 
     def create_and_test_config_common_properties(self):
         config = self.config_class(**self.inputs_dict)
-        self.parent.assertTrue(hasattr(config, "vocab_size"))
+        if self.has_text_modality:
+            self.parent.assertTrue(hasattr(config, "vocab_size"))
         self.parent.assertTrue(hasattr(config, "hidden_size"))
         self.parent.assertTrue(hasattr(config, "num_attention_heads"))
         self.parent.assertTrue(hasattr(config, "num_hidden_layers"))