Models doc (#7345)

* Clean up model documentation

* Formatting

* Preparation work

* Long lines

* Main work on rst files

* Cleanup all config files

* Syntax fix

* Clean all tokenizers

* Work on first models

* Models beginning

* FaluBERT

* All PyTorch models

* All models

* Long lines again

* Fixes

* More fixes

* Update docs/source/model_doc/bert.rst
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Update docs/source/model_doc/electra.rst
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Last fixes
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

docs/source/model_doc/pegasus.rst

 Pegasus
-----------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 **DISCLAIMER:** If you see something strange,
 file a `Github Issue <https://github.com/huggingface/transformers/issues/new?assignees=sshleifer&labels=&template=bug-report.md&title>`__ and assign
 @sshleifer.
 
 
 Overview
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The Pegasus model was proposed in `PEGASUS: Pre-training with Extracted Gap-sentences for
 Abstractive Summarization <https://arxiv.org/pdf/1912.08777.pdf>`_ by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu on Dec 18, 2019.
@@ -19,7 +19,7 @@ The Authors' code can be found `here <https://github.com/google-research/pegasus
 
 
 Checkpoints
-~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 All the `checkpoints <https://huggingface.co/models?search=pegasus>`_ are finetuned for summarization, besides ``pegasus-large``, whence the other checkpoints are finetuned.
 - Each checkpoint is 2.2 GB on disk and 568M parameters.
 - FP16 is not supported (help/ideas on this appreciated!).
@@ -29,7 +29,7 @@ The gap is likely because of different alpha/length_penalty implementations in b
 
 
 Implementation Notes
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 - All models are transformer encoder-decoders with 16 layers in each component.
 - The implementation is completely inherited from ``BartForConditionalGeneration``
@@ -43,7 +43,7 @@ Implementation Notes
 
 
 Usage Example
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. code-block:: python
 
@@ -63,7 +63,7 @@ Usage Example
     assert tgt_text[0] == "California's largest electricity provider has turned off power to hundreds of thousands of customers."
 
 PegasusForConditionalGeneration
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 This class inherits all functionality from ``BartForConditionalGeneration``, see that page for method signatures.
 Available models are listed at `Model List <https://huggingface.co/models?search=pegasus>`__
@@ -73,7 +73,7 @@ Available models are listed at `Model List <https://huggingface.co/models?search
 
 
 PegasusConfig
-~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 This config fully inherits from ``BartConfig``, but pegasus uses different default values:
 Up to date parameter values can be seen in `S3 <https://s3.amazonaws.com/models.huggingface.co/bert/google/pegasus-xsum/config.json>`_.
 As of Aug 10, 2020, they are:
@@ -107,7 +107,7 @@ As of Aug 10, 2020, they are:
 
 
 PegasusTokenizer
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 warning: ``add_tokens`` does not work at the moment.
 
 .. autoclass:: transformers.PegasusTokenizer

docs/source/model_doc/reformer.rst

 Reformer
-----------------------------------------------------
-**DISCLAIMER:** This model is still a work in progress, if you see something strange,
-file a `Github Issue <https://github.com/huggingface/transformers/issues/new?assignees=&labels=&template=bug-report.md&title>`_
+-----------------------------------------------------------------------------------------------------------------------
+
+**DISCLAIMER:** This model is still a work in progress, if you see something strange, file a `Github Issue
+<https://github.com/huggingface/transformers/issues/new?assignees=&labels=&template=bug-report.md&title>`__.
 
 Overview
-~~~~~~~~~~
-The Reformer model was presented in `Reformer: The Efficient Transformer <https://arxiv.org/abs/2001.04451.pdf>`_ by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-Here the abstract: 
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The Reformer model was proposed in the paper `Reformer: The Efficient Transformer
+<https://arxiv.org/abs/2001.04451.pdf>`__ by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
 
-*Large Transformer models routinely achieve state-of-the-art results on a number of tasks but training these models can be prohibitively costly, especially on long sequences. We introduce two techniques to improve the efficiency of Transformers. For one, we replace dot-product attention by one that uses locality-sensitive hashing, changing its complexity from O(L^2) to O(Llog(L)), where L is the length of the sequence. Furthermore, we use reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of N times, where N is the number of layers. The resulting model, the Reformer, performs on par with Transformer models while being much more memory-efficient and much faster on long sequences.*
+The abstract from the paper is the following: 
 
-The Authors' code can be found `here <https://github.com/google/trax/tree/master/trax/models/reformer>`_ .
+*Large Transformer models routinely achieve state-of-the-art results on a number of tasks but training these models can
+be prohibitively costly, especially on long sequences. We introduce two techniques to improve the efficiency of
+Transformers. For one, we replace dot-product attention by one that uses locality-sensitive hashing, changing its
+complexity from O(L^2) to O(Llog(L)), where L is the length of the sequence. Furthermore, we use reversible residual
+layers instead of the standard residuals, which allows storing activations only once in the training process instead of
+N times, where N is the number of layers. The resulting model, the Reformer, performs on par with Transformer models
+while being much more memory-efficient and much faster on long sequences.*
+
+The Authors' code can be found `here <https://github.com/google/trax/tree/master/trax/models/reformer>`__.
 
 Axial Positional Encodings
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-Axial Positional Encodings were first implemented in Google's `trax library <https://github.com/google/trax/blob/4d99ad4965bab1deba227539758d59f0df0fef48/trax/layers/research/position_encodings.py#L29>`_ and developed by the authors of this model's paper. In models that are treating very long input sequences, the conventional position id encodings store an embedings vector of size :math:`d` being the ``config.hidden_size`` for every position :math:`i, \ldots, n_s`, with :math:`n_s` being ``config.max_embedding_size``. *E.g.*, having a sequence length of :math:`n_s = 2^{19} \approx 0.5M` and a ``config.hidden_size`` of :math:`d = 2^{10} \approx 1000` would result in a position encoding matrix:
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Axial Positional Encodings were first implemented in Google's `trax library
+<https://github.com/google/trax/blob/4d99ad4965bab1deba227539758d59f0df0fef48/trax/layers/research/position_encodings.py#L29>`__
+and developed by the authors of this model's paper. In models that are treating very long input sequences, the
+conventional position id encodings store an embedings vector of size :math:`d` being the :obj:`config.hidden_size` for
+every position :math:`i, \ldots, n_s`, with :math:`n_s` being :obj:`config.max_embedding_size`. This means that having
+a sequence length of :math:`n_s = 2^{19} \approx 0.5M` and a ``config.hidden_size`` of :math:`d = 2^{10} \approx 1000`
+would result in a position encoding matrix:
 
 .. math::
     X_{i,j}, \text{ with } i \in \left[1,\ldots, d\right] \text{ and } j \in \left[1,\ldots, n_s\right] 
@@ -42,94 +59,127 @@ Therefore the following holds:
                 X^{2}_{i - d^1, l}, & \text{if } i \ge d^1 \text{ with } l = \lfloor\frac{j}{n_s^1}\rfloor
               \end{cases}
 
-Intuitively, this means that a position embedding vector :math:`x_j \in \mathbb{R}^{d}` is now the composition of two factorized embedding vectors: :math:`x^1_{k, l} + x^2_{l, k}`, where as the ``config.max_embedding_size`` dimension :math:`j` is factorized into :math:`k \text{ and } l`.
-This design ensures that each position embedding vector :math:`x_j` is unique.
+Intuitively, this means that a position embedding vector :math:`x_j \in \mathbb{R}^{d}` is now the composition of two
+factorized embedding vectors: :math:`x^1_{k, l} + x^2_{l, k}`, where as the :obj:`config.max_embedding_size` dimension
+:math:`j` is factorized into :math:`k \text{ and } l`. This design ensures that each position embedding vector
+:math:`x_j` is unique.
 
-Using the above example again, axial position encoding with :math:`d^1 = 2^5, d^2 = 2^5, n_s^1 = 2^9, n_s^2 = 2^{10}` can drastically reduced the number of parameters to :math:`2^{14} + 2^{15} \approx 49000` parameters.
-
-In practice, the parameter ``config.axial_pos_embds_dim`` is set to ``list``:math:`(d^1, d^2)` which sum has to be equal to ``config.hidden_size`` and ``config.axial_pos_shape`` is set to ``list``:math:`(n_s^1, n_s^2)` and which product has to be equal to ``config.max_embedding_size`` which during training has to be equal to the ``sequence length`` of the ``input_ids``.
+Using the above example again, axial position encoding with :math:`d^1 = 2^5, d^2 = 2^5, n_s^1 = 2^9, n_s^2 = 2^{10}`
+can drastically reduced the number of parameters to :math:`2^{14} + 2^{15} \approx 49000` parameters.
 
+In practice, the parameter :obj:`config.axial_pos_embds_dim` is set to a tuple :math:`(d^1, d^2)` which sum has to
+be equal to :obj:`config.hidden_size` and :obj:`config.axial_pos_shape` is set to a tuple :math:`(n_s^1, n_s^2)` which
+product has to be equal to :obj:`config.max_embedding_size`, which during training has to be equal to the
+`sequence length` of the :obj:`input_ids`.
 
 
 LSH Self Attention
-~~~~~~~~~~~~~~~~~~~~
-In Locality sensitive hashing (LSH) self attention the key and query projection weights are tied. Therefore, the key query embedding vectors are also tied.
-LSH self attention uses the locality sensitive 
-hashing mechanism proposed in `Practical and Optimal LSH for Angular Distance <https://arxiv.org/abs/1509.02897>`_ to assign each of the tied key query embedding vectors to one of ``config.num_buckets`` possible buckets. The premise is that the more "similar" key query embedding vectors (in terms of *cosine similarity*) are to each other, the more likely they are assigned to the same bucket. 
-The accuracy of the LSH mechanism can be improved by increasing ``config.num_hashes`` or directly the argument ``num_hashes`` of the forward function so that the output of the LSH self attention better approximates the output of the "normal" full self attention.
-The buckets are then sorted and chunked into query key embedding vector chunks each of length ``config.lsh_chunk_length``. For each chunk, the query embedding vectors attend to its key vectors (which are tied to themselves) and to the key embedding vectors of ``config.lsh_num_chunks_before`` previous neighboring chunks and ``config.lsh_num_chunks_after`` following neighboring chunks.
-For more information, see the `original Paper <https://arxiv.org/abs/2001.04451>`_ or this great `blog post <https://www.pragmatic.ml/reformer-deep-dive/>`_.
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+In Locality sensitive hashing (LSH) self attention the key and query projection weights are tied. Therefore, the key
+query embedding vectors are also tied. LSH self attention uses the locality sensitive hashing mechanism proposed in
+`Practical and Optimal LSH for Angular Distance <https://arxiv.org/abs/1509.02897>`__ to assign each of the tied key
+query embedding vectors to one of :obj:`config.num_buckets` possible buckets. The premise is that the more "similar"
+key query embedding vectors (in terms of *cosine similarity*) are to each other, the more likely they are assigned to
+the same bucket. 
+
+The accuracy of the LSH mechanism can be improved by increasing :obj:`config.num_hashes` or directly the argument 
+:obj:`num_hashes` of the forward function so that the output of the LSH self attention better approximates the output
+of the "normal" full self attention. The buckets are then sorted and chunked into query key embedding vector chunks
+each of length :obj:`config.lsh_chunk_length`. For each chunk, the query embedding vectors attend to its key vectors
+(which are tied to themselves) and to the key embedding vectors of :obj:`config.lsh_num_chunks_before` previous
+neighboring chunks and :obj:`config.lsh_num_chunks_after` following neighboring chunks.
+
+For more information, see the `original Paper <https://arxiv.org/abs/2001.04451>`__ or this great `blog post
+<https://www.pragmatic.ml/reformer-deep-dive/>`__.
+
+Note that :obj:`config.num_buckets` can also be factorized into a list
+:math:`(n_{\text{buckets}}^1, n_{\text{buckets}}^2)`. This way instead of assigning the query key embedding vectors to
+one of :math:`(1,\ldots, n_{\text{buckets}})` they are assigned to one of
+:math:`(1-1,\ldots, n_{\text{buckets}}^1-1, \ldots, 1-n_{\text{buckets}}^2, \ldots, n_{\text{buckets}}^1-n_{\text{buckets}}^2)`.
+This is crucial for very long sequences to save memory.
+
+When training a model from scratch, it is recommended to leave :obj:`config.num_buckets=None`, so that depending on the
+sequence length a good value for :obj:`num_buckets` is calculated on the fly. This value will then automatically be
+saved in the config and should be reused for inference.
+
+Using LSH self attention, the memory and time complexity of the query-key matmul operation can be reduced from
+:math:`\mathcal{O}(n_s \times n_s)` to :math:`\mathcal{O}(n_s \times \log(n_s))`, which usually represents the memory
+and time bottleneck in a transformer model, with :math:`n_s` being the sequence length.
 
-Note that ``config.num_buckets`` can also be factorized into a ``list``:math:`(n_{\text{buckets}}^1, n_{\text{buckets}}^2)`. This way instead of assigning the query key embedding vectors to one of :math:`(1,\ldots, n_{\text{buckets}})` they are assigned to one of :math:`(1-1,\ldots, n_{\text{buckets}}^1-1, \ldots, 1-n_{\text{buckets}}^2, \ldots, n_{\text{buckets}}^1-n_{\text{buckets}}^2)`. This is crucial for very long sequences to save memory.
 
-When training a model from scratch, it is recommended to leave ``config.num_buckets=None``, so that depending on the sequence length a good value for ``num_buckets`` is calculated on the fly. This value will then automatically be saved in the config and should be reused for inference.
+Local Self Attention
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-Using LSH self attention, the memory and time complexity of the query-key matmul operation can be reduced from :math:`\mathcal{O}(n_s \times n_s)` to :math:`\mathcal{O}(n_s \times \log(n_s))`, which usually represents the memory and time bottleneck in a transformer model, with :math:`n_s` being the sequence length.
+Local self attention is essentially a "normal" self attention layer with key, query and value projections, but is
+chunked so that in each chunk of length :obj:`config.local_chunk_length` the query embedding vectors only attends to
+the key embedding vectors in its chunk and to the key embedding vectors of :obj:`config.local_num_chunks_before`
+previous neighboring chunks and :obj:`config.local_num_chunks_after` following neighboring chunks.
 
+Using Local self attention, the memory and time complexity of the query-key matmul operation can be reduced from
+:math:`\mathcal{O}(n_s \times n_s)` to :math:`\mathcal{O}(n_s \times \log(n_s))`, which usually represents the memory
+and time bottleneck in a transformer model, with :math:`n_s` being the sequence length.
 
-Local Self Attention
-~~~~~~~~~~~~~~~~~~~~
-Local self attention is essentially a "normal" self attention layer with 
-key, query and value projections, but is chunked so that in each chunk of length ``config.local_chunk_length`` the query embedding vectors only attends to the key embedding vectors in its chunk and to the key embedding vectors of ``config.local_num_chunks_before`` previous neighboring chunks and ``config.local_num_chunks_after`` following neighboring chunks.
 
-Using Local self attention, the memory and time complexity of the query-key matmul operation can be reduced from :math:`\mathcal{O}(n_s \times n_s)` to :math:`\mathcal{O}(n_s \times \log(n_s))`, which usually represents the memory and time bottleneck in a transformer model, with :math:`n_s` being the sequence length.
+Training
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
+During training, we must ensure that the sequence length is set to a value that can be divided by the least common
+multiple of :obj:`config.lsh_chunk_length` and :obj:`config.local_chunk_length` and that the parameters of the Axial
+Positional Encodings are correctly set as described above. Reformer is very memory efficient so that the model can
+easily be trained on sequences as long as 64000 tokens.
 
-Training
-~~~~~~~~~~~~~~~~~~~~
-During training, we must ensure that the sequence length is set to a value that can be divided by the least common multiple of ``config.lsh_chunk_length`` and ``config.local_chunk_length`` and that the parameters of the Axial Positional Encodings are correctly set as described above. Reformer is very memory efficient so that the model can easily be trained on sequences as long as 64000 tokens.
-For training, the ``ReformerModelWithLMHead`` should be used as follows: 
+For training, the :class:`~transformers.ReformerModelWithLMHead` should be used as follows: 
 
-::
+.. code-block::
 
   input_ids = tokenizer.encode('This is a sentence from the training data', return_tensors='pt')
   loss = model(input_ids, labels=input_ids)[0]
 
 
 ReformerConfig
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerConfig
     :members:
 
 
 ReformerTokenizer
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerTokenizer
-    :members: 
+    :members: save_vocabulary
 
 
 ReformerModel
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerModel
-    :members:
+    :members: forward
 
 
 ReformerModelWithLMHead
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerModelWithLMHead
-    :members:
+    :members: forward
 
 
 ReformerForMaskedLM
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerForMaskedLM
-    :members:
+    :members: forward
 
 
 ReformerForSequenceClassification
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerForSequenceClassification
-    :members:
+    :members: forward
 
 
 ReformerForQuestionAnswering
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.ReformerForQuestionAnswering
-    :members:
+    :members: forward

docs/source/model_doc/retribert.rst

 RetriBERT
-----------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 
 Overview
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-The RetriBERT model was proposed in the blog post
-`Explain Anything Like I'm Five: A Model for Open Domain Long Form Question Answering <https://yjernite.github.io/lfqa.html>`__,
-RetriBERT is a small model that uses either a single or pair of Bert encoders with lower-dimension projection for dense semantic indexing of text.
+The RetriBERT model was proposed in the blog post `Explain Anything Like I'm Five: A Model for Open Domain Long Form
+Question Answering <https://yjernite.github.io/lfqa.html>`__. RetriBERT is a small model that uses either a single or
+pair of BERT encoders with lower-dimension projection for dense semantic indexing of text.
 
-Code to train and use the model can be found `here <https://github.com/huggingface/transformers/tree/master/examples/distillation>`_.
+Code to train and use the model can be found `here
+<https://github.com/huggingface/transformers/tree/master/examples/distillation>`__.
 
 
 RetriBertConfig
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RetriBertConfig
     :members:
 
 
 RetriBertTokenizer
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RetriBertTokenizer
     :members:
 
 
 RetriBertTokenizerFast
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RetriBertTokenizerFast
     :members:
 
 
 RetriBertModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RetriBertModel
-    :members:
+    :members: forward

docs/source/model_doc/roberta.rst

 RoBERTa
-----------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 
 Overview
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-The RoBERTa model was proposed in `RoBERTa: A Robustly Optimized BERT Pretraining Approach <https://arxiv.org/abs/1907.11692>`_
-by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer,
-Veselin Stoyanov. It is based on Google's BERT model released in 2018.
+The RoBERTa model was proposed in `RoBERTa: A Robustly Optimized BERT Pretraining Approach
+<https://arxiv.org/abs/1907.11692>`_ by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer
+Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov. It is based on Google's BERT model released in 2018.
 
 It builds on BERT and modifies key hyperparameters, removing the next-sentence pretraining
 objective and training with much larger mini-batches and learning rates.
@@ -27,22 +27,23 @@ Tips:
 - This implementation is the same as :class:`~transformers.BertModel` with a tiny embeddings tweak as well as a
   setup for Roberta pretrained models.
 - RoBERTa has the same architecture as BERT, but uses a byte-level BPE as a tokenizer (same as GPT-2) and uses a
-  different pre-training scheme.
-- RoBERTa doesn't have `token_type_ids`, you don't need to indicate which token belongs to which segment. Just separate your segments with the separation token `tokenizer.sep_token` (or `</s>`)
-- `Camembert <./camembert.html>`__ is a wrapper around RoBERTa. Refer to this page for usage examples.
+  different pretraining scheme.
+- RoBERTa doesn't have :obj:`token_type_ids`, you don't need to indicate which token belongs to which segment. Just
+  separate your segments with the separation token :obj:`tokenizer.sep_token` (or :obj:`</s>`)
+- :doc:`CamemBERT <camembert>` is a wrapper around RoBERTa. Refer to this page for usage examples.
 
 The original code can be found `here <https://github.com/pytorch/fairseq/tree/master/examples/roberta>`_.
 
 
 RobertaConfig
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaConfig
     :members:
 
 
 RobertaTokenizer
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaTokenizer
     :members: build_inputs_with_special_tokens, get_special_tokens_mask,
@@ -50,98 +51,98 @@ RobertaTokenizer
 
 
 RobertaTokenizerFast
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaTokenizerFast
     :members: build_inputs_with_special_tokens
 
 
 RobertaModel
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaModel
-    :members:
+    :members: forward
 
 
 RobertaForCausalLM
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForCausalLM
-    :members:
+    :members: forward
 
 
 RobertaForMaskedLM
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForMaskedLM
-    :members:
+    :members: forward
 
 
 RobertaForSequenceClassification
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForSequenceClassification
-    :members:
+    :members: forward
 
 
 RobertaForMultipleChoice
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForMultipleChoice
-    :members:
+    :members: forward
 
 
 RobertaForTokenClassification
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForTokenClassification
-    :members:
+    :members: forward
 
 
 RobertaForQuestionAnswering
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.RobertaForQuestionAnswering
-    :members:
+    :members: forward
 
 
 TFRobertaModel
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaModel
-    :members:
+    :members: call
 
 
 TFRobertaForMaskedLM
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaForMaskedLM
-    :members:
+    :members: call
 
 
 TFRobertaForSequenceClassification
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaForSequenceClassification
-    :members:
+    :members: call
 
 
 TFRobertaForMultipleChoice
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaForMultipleChoice
-    :members:
+    :members: call
 
 
 TFRobertaForTokenClassification
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaForTokenClassification
-    :members:
+    :members: call
 
 
 TFRobertaForQuestionAnswering
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFRobertaForQuestionAnswering
-    :members:
+    :members: call

docs/source/model_doc/t5.rst

 T5
-----------------------------------------------------
-**DISCLAIMER:** This model is still a work in progress, if you see something strange,
-file a `Github Issue <https://github.com/huggingface/transformers/issues/new?assignees=&labels=&template=bug-report.md&title>`_
+-----------------------------------------------------------------------------------------------------------------------
+
+**DISCLAIMER:** This model is still a work in progress, if you see something strange, file a `Github Issue
+<https://github.com/huggingface/transformers/issues/new?assignees=&labels=&template=bug-report.md&title>`__.
 
 Overview
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The T5 model was presented in `Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
+<https://arxiv.org/pdf/1910.10683.pdf>`_ by Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang,
+Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu.
 
-The T5 model was presented in `Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer <https://arxiv.org/pdf/1910.10683.pdf>`_ by Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu in 
-Here the abstract: 
+The abstract from the paper is the following:
 
-*Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. 
-In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. 
-Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. 
-By combining the insights from our exploration with scale and our new "Colossal Clean Crawled Corpus", we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. 
-To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.*
+*Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream
+task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning
+has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of
+transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a
+text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer
+approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration
+with scale and our new "Colossal Clean Crawled Corpus", we achieve state-of-the-art results on many benchmarks covering
+summarization, question answering, text classification, and more. To facilitate future work on transfer learning for
+NLP, we release our dataset, pre-trained models, and code.*
 
 Tips:
 
-- T5 is an encoder-decoder model pre-trained on a multi-task mixture of unsupervised 
-  and supervised tasks and for which each task is converted into a text-to-text format.
-  T5 works well on a variety of tasks out-of-the-box by prepending a different prefix to the input corresponding to each task, e.g.: for translation: *translate English to German: ..., summarize: ...*.
-  For more information about which prefix to use, it is easiest to look into Appendix D of the `paper <https://arxiv.org/pdf/1910.10683.pdf>`_ .
-- For sequence to sequence generation, it is recommended to use ``T5ForConditionalGeneration.generate()``. The method takes care of feeding the encoded input via cross-attention layers to the decoder and auto-regressively generates the decoder output.
+- T5 is an encoder-decoder model pre-trained on a multi-task mixture of unsupervised and supervised tasks and for which
+  each task is converted into a text-to-text format. T5 works well on a variety of tasks out-of-the-box by prepending a
+  different prefix to the input corresponding to each task, e.g., for translation: *translate English to German: ...*,
+  for summarization: *summarize: ...*.
+  
+  For more information about which prefix to use, it is easiest to look into Appendix D of the `paper
+  <https://arxiv.org/pdf/1910.10683.pdf>`__.
+- For sequence-to-sequence generation, it is recommended to use :obj:`T5ForConditionalGeneration.generate()``. This
+  method takes care of feeding the encoded input via cross-attention layers to the decoder and auto-regressively
+  generates the decoder output.
 - T5 uses relative scalar embeddings. Encoder input padding can be done on the left and on the right.
 
-The original code can be found `here <https://github.com/google-research/text-to-text-transfer-transformer>`_.
+The original code can be found `here <https://github.com/google-research/text-to-text-transfer-transformer>`__.
 
 Training
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-T5 is an encoder-decoder model and converts all NLP problems into a text-to-text format. It is trained using teacher forcing.
-This means that for training we always need an input sequence and a target sequence. 
-The input sequence is fed to the model using ``input_ids``. The target sequence is shifted to the right, *i.e.* prepended by a start-sequence token and fed to the decoder using the `decoder_input_ids`. In teacher-forcing style, the target sequence is then appended by the EOS token and corresponds to the ``labels``. The PAD token is hereby used as the start-sequence token.
-T5 can be trained / fine-tuned both in a supervised and unsupervised fashion.
+T5 is an encoder-decoder model and converts all NLP problems into a text-to-text format. It is trained using teacher
+forcing. This means that for training we always need an input sequence and a target sequence. The input sequence is fed
+to the model using :obj:`input_ids``. The target sequence is shifted to the right, i.e., prepended by a start-sequence
+token and fed to the decoder using the :obj:`decoder_input_ids`. In teacher-forcing style, the target sequence is then
+appended by the EOS token and corresponds to the :obj:`labels`. The PAD token is hereby used as the start-sequence
+token. T5 can be trained / fine-tuned both in a supervised and unsupervised fashion.
 
 - Unsupervised denoising training
 
   In this setup spans of the input sequence are masked by so-called sentinel tokens (*a.k.a* unique mask tokens) 
   and the output sequence is formed as a concatenation of the same sentinel tokens and the *real* masked tokens. 
-  Each sentinel token represents a unique mask token for this sentence and should start with ``<extra_id_0>``, ``<extra_id_1>``, ... up to ``<extra_id_99>``. As a default 100 sentinel tokens are available in ``T5Tokenizer``.
-  *E.g.* the sentence "The cute dog walks in the park" with the masks put on "cute dog" and "the" should be processed as follows: 
+  Each sentinel token represents a unique mask token for this sentence and should start with :obj:`<extra_id_0>`, 
+  :obj:`<extra_id_1>`, ... up to :obj:`<extra_id_99>`. As a default, 100 sentinel tokens are available in
+  :class:`~transformers.T5Tokenizer`.
+  
+  For instance, the sentence "The cute dog walks in the park" with the masks put on "cute dog" and "the" should be
+  processed as follows: 
 
-::
+.. code-block::
 
   input_ids = tokenizer.encode('The <extra_id_0> walks in <extra_id_1> park', return_tensors='pt')
   labels = tokenizer.encode('<extra_id_0> cute dog <extra_id_1> the <extra_id_2> </s>', return_tensors='pt')
@@ -50,11 +69,11 @@ T5 can be trained / fine-tuned both in a supervised and unsupervised fashion.
 
 - Supervised training
 
-  In this setup the input sequence and output sequence are standard sequence to sequence input output mapping.
-  In translation, *e.g.* the input sequence "The house is wonderful." and output sequence "Das Haus ist wunderbar." should 
-  be processed as follows:
+  In this setup the input sequence and output sequence are standard sequence-to-sequence input output mapping.
+  In translation, for instance with the input sequence "The house is wonderful." and output sequence "Das Haus ist
+  wunderbar.", the sentences should be processed as follows:
   
-::
+.. code-block::
 
   input_ids = tokenizer.encode('translate English to German: The house is wonderful. </s>', return_tensors='pt')
   labels = tokenizer.encode('Das Haus ist wunderbar. </s>', return_tensors='pt')
@@ -63,43 +82,43 @@ T5 can be trained / fine-tuned both in a supervised and unsupervised fashion.
 
 
 T5Config
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.T5Config
     :members:
 
 
 T5Tokenizer
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.T5Tokenizer
     :members: build_inputs_with_special_tokens, get_special_tokens_mask,
-        create_token_type_ids_from_sequences, save_vocabulary
+        create_token_type_ids_from_sequences, prepare_seq2seq_batch, save_vocabulary
 
 
 T5Model
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.T5Model
-    :members:
+    :members: forward
 
 
 T5ForConditionalGeneration
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.T5ForConditionalGeneration
-    :members:
+    :members: forward
 
 
 TFT5Model
-~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFT5Model
-    :members:
+    :members: call
 
 
 TFT5ForConditionalGeneration
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFT5ForConditionalGeneration
-    :members:
+    :members: call

docs/source/model_doc/transformerxl.rst

 Transformer XL
-----------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 
 Overview
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-The Transformer-XL model was proposed in
-`Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context <https://arxiv.org/abs/1901.02860>`__
-by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
-It's a causal (uni-directional) transformer with relative positioning (sinusoïdal) embeddings which can reuse
-previously computed hidden-states to attend to longer context (memory).
-This model also uses adaptive softmax inputs and outputs (tied).
+The Transformer-XL model was proposed in `Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
+<https://arxiv.org/abs/1901.02860>`__ by Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan
+Salakhutdinov. It's a causal (uni-directional) transformer with relative positioning (sinusoïdal) embeddings which can
+reuse previously computed hidden-states to attend to longer context (memory). This model also uses adaptive softmax
+inputs and outputs (tied).
 
 The abstract from the paper is the following:
 
@@ -30,32 +29,32 @@ Tips:
   The original implementation trains on SQuAD with padding on the left, therefore the padding defaults are set to left.
 - Transformer-XL is one of the few models that has no sequence length limit.
 
-The original code can be found `here <https://github.com/kimiyoung/transformer-xl>`_.
+The original code can be found `here <https://github.com/kimiyoung/transformer-xl>`__.
 
 
 TransfoXLConfig
-~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TransfoXLConfig
     :members:
 
 
 TransfoXLTokenizer
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TransfoXLTokenizer
     :members: save_vocabulary
 
 
 TransfoXLTokenizerFast
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TransfoXLTokenizerFast
     :members:
 
 
 TransfoXL specific outputs
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.modeling_transfo_xl.TransfoXLModelOutput
     :members:
@@ -71,28 +70,28 @@ TransfoXL specific outputs
 
 
 TransfoXLModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TransfoXLModel
-    :members:
+    :members: forward
 
 
 TransfoXLLMHeadModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TransfoXLLMHeadModel
-    :members:
+    :members: forward
 
 
 TFTransfoXLModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFTransfoXLModel
-    :members:
+    :members: call
 
 
 TFTransfoXLLMHeadModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: transformers.TFTransfoXLLMHeadModel
-    :members:
+    :members: call

docs/source/multilingual.rst

 Multi-lingual models
-================================================
+=======================================================================================================================
 
 Most of the models available in this library are mono-lingual models (English, Chinese and German). A few
 multi-lingual models are available and have a different mechanisms than mono-lingual models.
@@ -8,13 +8,13 @@ This page details the usage of these models.
 The two models that currently support multiple languages are BERT and XLM.
 
 XLM
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 XLM has a total of 10 different checkpoints, only one of which is mono-lingual. The 9 remaining model checkpoints can
 be split in two categories: the checkpoints that make use of language embeddings, and those that don't
 
 XLM & Language Embeddings
-------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 
 This section concerns the following checkpoints:
 
@@ -82,7 +82,7 @@ The example `run_generation.py <https://github.com/huggingface/transformers/blob
 can generate text using the CLM checkpoints from XLM, using the language embeddings.
 
 XLM without Language Embeddings
-------------------------------------------------
+-----------------------------------------------------------------------------------------------------------------------
 
 This section concerns the following checkpoints:
 
@@ -94,7 +94,7 @@ sentence representations, differently from previously-mentioned XLM checkpoints.
 
 
 BERT
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 BERT has two checkpoints that can be used for multi-lingual tasks:
 
@@ -105,7 +105,7 @@ These checkpoints do not require language embeddings at inference time. They sho
 used in the context and infer accordingly.
 
 XLM-RoBERTa
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 XLM-RoBERTa was trained on 2.5TB of newly created clean CommonCrawl data in 100 languages. It provides strong
 gains over previously released multi-lingual models like mBERT or XLM on downstream taks like classification,

docs/source/perplexity.rst

 Perplexity of fixed-length models
-=================================
+=======================================================================================================================
 
 Perplexity (PPL) is one of the most common metrics for evaluating language
 models. Before diving in, we should note that the metric applies specifically
@@ -31,7 +31,7 @@ relationship to Bits Per Character (BPC) and data compression, check out this
 <https://thegradient.pub/understanding-evaluation-metrics-for-language-models/>`_.
 
 Calculating PPL with fixed-length models
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 If we weren't limited by a model's context size, we would evaluate the
 model's perplexity by autoregressively factorizing a sequence and
@@ -83,7 +83,7 @@ time. This allows computation to procede much faster while still giving the
 model a large context to make predictions at each step.
 
 Example: Calculating perplexity with GPT-2 in 🤗 Transformers
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Let's demonstrate this process with GPT-2.
 

docs/source/philosophy.rst

 Philosophy
-==========
+=======================================================================================================================
 
 🤗 Transformers is an opinionated library built for:
 
@@ -48,7 +48,7 @@ A few other goals:
 - Switch easily between PyTorch and TensorFlow 2.0, allowing training using one framework and inference using another.
 
 Main concepts
-~~~~~~~~~~~~~
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The library is built around three types of classes for each model:
 

docs/source/pretrained_models.rst

 Pretrained models
-================================================
+=======================================================================================================================
 
 Here is the full list of the currently provided pretrained models together with a short presentation of each model.