[doc] fix broken ref (#12597)

src/transformers/models/xlm_prophetnet/tokenization_xlm_prophetnet.py

@@ -56,7 +56,7 @@ def load_vocab(vocab_file):
 
 class XLMProphetNetTokenizer(PreTrainedTokenizer):
     """
-    Adapted from :class:`~transformers.RobertaTokenizer` and class:`~transformers.XLNetTokenizer`. Based on
+    Adapted from :class:`~transformers.RobertaTokenizer` and :class:`~transformers.XLNetTokenizer`. Based on
     `SentencePiece <https://github.com/google/sentencepiece>`__.
 
     This tokenizer inherits from :class:`~transformers.PreTrainedTokenizer` which contains most of the main methods.

src/transformers/models/xlm_roberta/tokenization_xlm_roberta.py

@@ -54,7 +54,7 @@ PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
 
 class XLMRobertaTokenizer(PreTrainedTokenizer):
     """
-    Adapted from :class:`~transformers.RobertaTokenizer` and class:`~transformers.XLNetTokenizer`. Based on
+    Adapted from :class:`~transformers.RobertaTokenizer` and :class:`~transformers.XLNetTokenizer`. Based on
     `SentencePiece <https://github.com/google/sentencepiece>`__.
 
     This tokenizer inherits from :class:`~transformers.PreTrainedTokenizer` which contains most of the main methods.

src/transformers/models/xlm_roberta/tokenization_xlm_roberta_fast.py

@@ -67,7 +67,7 @@ PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
 class XLMRobertaTokenizerFast(PreTrainedTokenizerFast):
     """
     Construct a "fast" XLM-RoBERTa tokenizer (backed by HuggingFace's `tokenizers` library). Adapted from
-    :class:`~transformers.RobertaTokenizer` and class:`~transformers.XLNetTokenizer`. Based on `BPE
+    :class:`~transformers.RobertaTokenizer` and :class:`~transformers.XLNetTokenizer`. Based on `BPE
     <https://huggingface.co/docs/tokenizers/python/latest/components.html?highlight=BPE#models>`__.
 
     This tokenizer inherits from :class:`~transformers.PreTrainedTokenizerFast` which contains most of the main