Add FastTokenizer to REALM (#15211)

* Remove BertTokenizer abstraction

* Add FastTokenizer to REALM

* Fix config archive map

* Fix copies

* Update realm.mdx

* Apply suggestions from code review

docs/source/index.mdx

@@ -246,7 +246,7 @@ Flax), PyTorch, and/or TensorFlow.
 |         ProphetNet          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           QDQBert           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             RAG             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            Realm            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            Realm            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          Reformer           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |           RemBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |          RetriBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |

docs/source/model_doc/realm.mdx

@@ -49,6 +49,11 @@ This model was contributed by [qqaatw](https://huggingface.co/qqaatw). The origi
     - save_vocabulary
     - batch_encode_candidates
 
+## RealmTokenizerFast
+
+[[autodoc]] RealmTokenizerFast
+    - batch_encode_candidates
+
 ## RealmRetriever
 
 [[autodoc]] RealmRetriever