[PyTorch Bart] Split Bart into different models (#9343)

* first try

* remove old template

* finish bart

* finish mbart

* delete unnecessary line

* init pegasus

* save intermediate

* correct pegasus

* finish pegasus

* remove cookie cutter leftover

* add marian

* finish blenderbot

* replace in file

* correctly split blenderbot

* delete "old" folder

* correct "add statement"

* adapt config for tf comp

* correct configs for tf

* remove ipdb

* fix more stuff

* fix mbart

* push pegasus fix

* fix mbart

* more fixes

* fix research projects code

* finish docs for bart, mbart, and marian

* delete unnecessary file

* correct attn typo

* correct configs

* remove pegasus for seq class

* correct peg docs

* correct peg docs

* finish configs

* further improve docs

* add copied from statements to mbart

* fix copied from in mbart

* add copy statements to marian

* add copied from to marian

* add pegasus copied from

* finish pegasus

* finish copied from

* Apply suggestions from code review

* make style

* backward comp blenderbot

* apply lysandres and sylvains suggestions

* apply suggestions

* push last fixes

* fix docs

* fix tok tests

* fix imports code style

* fix doc

src/transformers/models/t5/modeling_t5.py

@@ -1028,14 +1028,22 @@ T5_INPUTS_DOCSTRING = r"""
 
             `What are attention masks? <../glossary.html#attention-mask>`__
         decoder_input_ids (:obj:`torch.LongTensor` of shape :obj:`(batch_size, target_sequence_length)`, `optional`):
-            Provide for sequence to sequence training. T5 uses the :obj:`pad_token_id` as the starting token for
-            :obj:`decoder_input_ids` generation. If :obj:`past_key_values` is used, optionally only the last
-            :obj:`decoder_input_ids` have to be input (see :obj:`past_key_values`).
+            Indices of decoder input sequence tokens in the vocabulary.
+
+            Indices can be obtained using :class:`~transformers.BartTokenizer`. See
+            :meth:`transformers.PreTrainedTokenizer.encode` and :meth:`transformers.PreTrainedTokenizer.__call__` for
+            details.
+
+            `What are input IDs? <../glossary.html#input-ids>`__
+
+            T5 uses the :obj:`pad_token_id` as the starting token for :obj:`decoder_input_ids` generation. If
+            :obj:`past_key_values` is used, optionally only the last :obj:`decoder_input_ids` have to be input (see
+            :obj:`past_key_values`).
 
             To know more on how to prepare :obj:`decoder_input_ids` for pretraining take a look at `T5 Training
             <./t5.html#training>`__. If :obj:`decoder_input_ids` and :obj:`decoder_inputs_embeds` are both unset,
             :obj:`decoder_input_ids` takes the value of :obj:`input_ids`.
-        decoder_attention_mask (:obj:`torch.BoolTensor` of shape :obj:`(batch_size, tgt_seq_len)`, `optional`):
+        decoder_attention_mask (:obj:`torch.BoolTensor` of shape :obj:`(batch_size, target_sequence_length)`, `optional`):
             Default behavior: generate a tensor that ignores pad tokens in :obj:`decoder_input_ids`. Causal mask will
             also be used by default.
         encoder_outputs (:obj:`tuple(tuple(torch.FloatTensor)`, `optional`):

src/transformers/models/t5/modeling_tf_t5.py

@@ -922,7 +922,7 @@ T5_INPUTS_DOCSTRING = r"""
             - 0 for tokens that are **masked**.
 
             `What are attention masks? <../glossary.html#attention-mask>`__
-        decoder_attention_mask (:obj:`tf.Tensor` of shape :obj:`(batch_size, tgt_seq_len)`, `optional`):
+        decoder_attention_mask (:obj:`tf.Tensor` of shape :obj:`(batch_size, target_sequence_length)`, `optional`):
             Default behavior: generate a tensor that ignores pad tokens in :obj:`decoder_input_ids`. Causal mask will
             also be used by default.
         encoder_outputs (:obj:`tuple(tuple(tf.FloatTensor)`, `optional`):

src/transformers/utils/dummy_pt_objects.py

@@ -609,6 +609,27 @@ class BlenderbotModel:
         requires_pytorch(self)
 
 
+BLENDERBOT_SMALL_PRETRAINED_MODEL_ARCHIVE_LIST = None
+
+
+class BlenderbotSmallForConditionalGeneration:
+    def __init__(self, *args, **kwargs):
+        requires_pytorch(self)
+
+    @classmethod
+    def from_pretrained(self, *args, **kwargs):
+        requires_pytorch(self)
+
+
+class BlenderbotSmallModel:
+    def __init__(self, *args, **kwargs):
+        requires_pytorch(self)
+
+    @classmethod
+    def from_pretrained(self, *args, **kwargs):
+        requires_pytorch(self)
+
+
 CAMEMBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -1327,6 +1348,15 @@ class LxmertXLayer:
         requires_pytorch(self)
 
 
+class MarianModel:
+    def __init__(self, *args, **kwargs):
+        requires_pytorch(self)
+
+    @classmethod
+    def from_pretrained(self, *args, **kwargs):
+        requires_pytorch(self)
+
+
 class MarianMTModel:
     def __init__(self, *args, **kwargs):
         requires_pytorch(self)
@@ -1345,6 +1375,24 @@ class MBartForConditionalGeneration:
         requires_pytorch(self)
 
 
+class MBartForQuestionAnswering:
+    def __init__(self, *args, **kwargs):
+        requires_pytorch(self)
+
+    @classmethod
+    def from_pretrained(self, *args, **kwargs):
+        requires_pytorch(self)
+
+
+class MBartForSequenceClassification:
+    def __init__(self, *args, **kwargs):
+        requires_pytorch(self)
+
+    @classmethod
+    def from_pretrained(self, *args, **kwargs):
+        requires_pytorch(self)
+
+
 class MBartModel:
     def __init__(self, *args, **kwargs):
         requires_pytorch(self)

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/configuration_{{cookiecutter.lowercase_modelname}}.py

@@ -135,6 +135,10 @@ class {{cookiecutter.camelcase_modelname}}Config(PretrainedConfig):
         >>> configuration = model.config
     """
     model_type = "{{cookiecutter.lowercase_modelname}}"
+    {% if cookiecutter.is_encoder_decoder_model == "False" -%}
+    {% else -%}
+    keys_to_ignore_at_inference = ["past_key_values"]
+    {% endif -%}
 
     def __init__(
         self,

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py

@@ -1849,7 +1849,7 @@ class TF{{cookiecutter.camelcase_modelname}}PreTrainedModel(TFPreTrainedModel):
         decoder_input_ids (:obj:`tf.Tensor` of shape :obj:`(batch_size, target_sequence_length)`, `optional`):
             Provide for translation and summarization training. By default, the model will create this tensor by
             shifting the input_ids right, following the paper.
-        decoder_attention_mask (:obj:`tf.Tensor` of shape :obj:`(batch_size, tgt_seq_len)`, `optional`):
+        decoder_attention_mask (:obj:`tf.Tensor` of shape :obj:`(batch_size, target_sequence_length)`, `optional`):
             will be made by default and ignore pad tokens. It is not recommended to set this for most use cases.
         encoder_outputs (:obj:`tf.FloatTensor`, `optional`):
             hidden states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/test_modeling_{{cookiecutter.lowercase_modelname}}.py

@@ -480,8 +480,6 @@ import copy
 import tempfile
 import unittest
 
-import timeout_decorator  # noqa
-
 from transformers import is_torch_available
 from transformers.file_utils import cached_property
 from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow, torch_device