Fix TF start docstrings (#18991)

* Update our TF 2.0 input format tip across all models

* make style

src/transformers/models/albert/modeling_tf_albert.py

@@ -694,23 +694,28 @@ ALBERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/bart/modeling_tf_bart.py

@@ -514,16 +514,17 @@ BART_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
-
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
     - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
@@ -531,6 +532,10 @@ BART_START_DOCSTRING = r"""
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/bert/modeling_tf_bert.py

@@ -964,23 +964,28 @@ BERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/blenderbot/modeling_tf_blenderbot.py

@@ -514,16 +514,17 @@ BLENDERBOT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
-
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
     - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
@@ -531,6 +532,10 @@ BLENDERBOT_START_DOCSTRING = r"""
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/blenderbot_small/modeling_tf_blenderbot_small.py

@@ -514,16 +514,17 @@ BLENDERBOT_SMALL_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
-
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
     - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
@@ -531,6 +532,10 @@ BLENDERBOT_SMALL_START_DOCSTRING = r"""
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/camembert/modeling_tf_camembert.py

@@ -47,23 +47,28 @@ CAMEMBERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/clip/modeling_tf_clip.py

@@ -913,22 +913,27 @@ CLIP_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
-
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
     - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
-      `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
+    `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
-      `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
+    `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
+
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
 
     </Tip>
 

src/transformers/models/convbert/modeling_tf_convbert.py

@@ -638,23 +638,28 @@ CONVBERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Args:

src/transformers/models/convnext/modeling_tf_convnext.py

@@ -399,13 +399,27 @@ CONVNEXT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
-
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    - having all inputs as a list, tuple or dict in the first positional argument.
+
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
+
+    - a single Tensor with `pixel_values` only and nothing else: `model(pixel_values)`
+    - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
+    `model([pixel_values, attention_mask])` or `model([pixel_values, attention_mask, token_type_ids])`
+    - a dictionary with one or several input Tensors associated to the input names given in the docstring:
+    `model({"pixel_values": pixel_values, "token_type_ids": token_type_ids})`
+
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
 
     </Tip>
 

src/transformers/models/ctrl/modeling_tf_ctrl.py

@@ -418,23 +418,28 @@ CTRL_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/data2vec/modeling_tf_data2vec_vision.py

@@ -816,13 +816,27 @@ DATA2VEC_VISION_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
-
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    - having all inputs as a list, tuple or dict in the first positional argument.
+
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
+
+    - a single Tensor with `pixel_values` only and nothing else: `model(pixel_values)`
+    - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
+    `model([pixel_values, attention_mask])` or `model([pixel_values, attention_mask, token_type_ids])`
+    - a dictionary with one or several input Tensors associated to the input names given in the docstring:
+    `model({"pixel_values": pixel_values, "token_type_ids": token_type_ids})`
+
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
 
     </Tip>
 

src/transformers/models/deberta/modeling_tf_deberta.py

@@ -1004,23 +1004,28 @@ DEBERTA_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/deberta_v2/modeling_tf_deberta_v2.py

@@ -1097,23 +1097,28 @@ DEBERTA_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/distilbert/modeling_tf_distilbert.py

@@ -447,22 +447,27 @@ DISTILBERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
-    `model([input_ids, attention_mask])`
+    `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
-    `model({"input_ids": input_ids})`
+    `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
+
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
 
     </Tip>
 

src/transformers/models/dpr/modeling_tf_dpr.py

@@ -404,23 +404,28 @@ TF_DPR_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/electra/modeling_tf_electra.py

@@ -841,23 +841,28 @@ ELECTRA_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/flaubert/modeling_tf_flaubert.py

@@ -73,23 +73,28 @@ FLAUBERT_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/funnel/modeling_tf_funnel.py

@@ -1015,23 +1015,28 @@ FUNNEL_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/gpt2/modeling_tf_gpt2.py

@@ -607,23 +607,28 @@ GPT2_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters:

src/transformers/models/gptj/modeling_tf_gptj.py

@@ -550,23 +550,28 @@ GPTJ_START_DOCSTRING = r"""
 
     <Tip>
 
-    TF 2.0 models accepts two formats as inputs:
+    TensorFlow models and layers in `transformers` accept two formats as input:
 
     - having all inputs as keyword arguments (like PyTorch models), or
-    - having all inputs as a list, tuple or dict in the first positional arguments.
+    - having all inputs as a list, tuple or dict in the first positional argument.
 
-    This second option is useful when using [`tf.keras.Model.fit`] method which currently requires having all the
-    tensors in the first argument of the model call function: `model(inputs)`.
+    The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
+    and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
+    pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second
+    format outside of Keras methods like `fit()` and `predict()`, such as when creating your own layers or models with
+    the Keras `Functional` API, there are three possibilities you can use to gather all the input Tensors in the first
+    positional argument:
 
-    If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the
-    first positional argument :
-
-    - a single Tensor with `input_ids` only and nothing else: `model(inputs_ids)`
+    - a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
     - a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
     `model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
     - a dictionary with one or several input Tensors associated to the input names given in the docstring:
     `model({"input_ids": input_ids, "token_type_ids": token_type_ids})`
 
+    Note that when creating models and layers with
+    [subclassing](https://keras.io/guides/making_new_layers_and_models_via_subclassing/) then you don't need to worry
+    about any of this, as you can just pass inputs like you would to any other Python function!
+
     </Tip>
 
     Parameters: