Add barthez model (#8393)

* Add init barthez

* Add barthez model, tokenizer and docs

BARThez is a pre-trained french seq2seq model that uses BART objective.

* Apply suggestions from code review docs typos

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Add license

* Change URLs scheme

* Remove barthez model keep tokenizer

* Fix style

* Fix quality

* Update tokenizer

* Add fast tokenizer

* Add fast tokenizer test

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

docs/source/model_doc/barthez.rst

@@ -0,0 +1,41 @@
+BARThez
+-----------------------------------------------------------------------------------------------------------------------
+
+Overview
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The BARThez model was proposed in `BARThez: a Skilled Pretrained French Sequence-to-Sequence Model`
+<https://arxiv.org/abs/2010.12321>`__ by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis on 23 Oct,
+2020.
+
+The abstract of the paper:
+
+
+*Inductive transfer learning, enabled by self-supervised learning, have taken the entire Natural Language Processing
+(NLP) field by storm, with models such as BERT and BART setting new state of the art on countless natural language
+understanding tasks. While there are some notable exceptions, most of the available models and research have been
+conducted for the English language. In this work, we introduce BARThez, the first BART model for the French language
+(to the best of our knowledge). BARThez was pretrained on a very large monolingual French corpus from past research
+that we adapted to suit BART's perturbation schemes. Unlike already existing BERT-based French language models such as
+CamemBERT and FlauBERT, BARThez is particularly well-suited for generative tasks, since not only its encoder but also
+its decoder is pretrained. In addition to discriminative tasks from the FLUE benchmark, we evaluate BARThez on a novel
+summarization dataset, OrangeSum, that we release with this paper. We also continue the pretraining of an already
+pretrained multilingual BART on BARThez's corpus, and we show that the resulting model, which we call mBARTHez,
+provides a significant boost over vanilla BARThez, and is on par with or outperforms CamemBERT and FlauBERT.*
+
+The Authors' code can be found `here <https://github.com/moussaKam/BARThez>`__.
+
+
+Examples
+_______________________________________________________________________________________________________________________
+
+- BARThez can be fine-tuned on sequence-to-sequence tasks in a similar way as BART, check: `examples/seq2seq/
+  <https://github.com/huggingface/transformers/blob/master/examples/seq2seq/README.md>`__.
+
+
+BarthezTokenizer
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: transformers.BarthezTokenizer
+    :members:
+