[official/**.py] Rename "Arguments:" to "Args:"

official/nlp/modeling/layers/tn_expand_condense.py

@@ -36,7 +36,7 @@ class TNExpandCondense(Layer):
 
   Note the input shape and output shape will be identical.
 
-  Arguments:
+  Args:
     proj_multiplier: Positive integer, multiple of input_shape[-1] to project
       up to. Must be one of [2, 4, 6, 8].
     use_bias: Boolean, whether the layer uses a bias vector.

official/nlp/modeling/layers/tn_transformer_expand_condense.py

@@ -31,7 +31,7 @@ class TNTransformerExpandCondense(tf.keras.layers.Layer):
   tensor network layer replacing the usual intermediate and output Dense
   layers.
 
-  Arguments:
+  Args:
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate layer.
     intermediate_activation: Activation for the intermediate layer.

official/nlp/modeling/layers/transformer.py

@@ -31,7 +31,7 @@ class Transformer(keras_nlp.layers.TransformerEncoderBlock):
   This layer implements the Transformer from "Attention Is All You Need".
   (https://arxiv.org/abs/1706.03762).
 
-  Arguments:
+  Args:
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate layer.
     intermediate_activation: Activation for the intermediate layer.
@@ -117,7 +117,7 @@ class TransformerDecoderBlock(tf.keras.layers.Layer):
   (2) a encoder-decoder attention.
   (3) a positionwise fully connected feed-forward network.
 
-  Arguments:
+  Args:
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate layer.
     intermediate_activation: Activation for the intermediate layer.

official/nlp/modeling/layers/transformer_scaffold.py

@@ -35,7 +35,7 @@ class TransformerScaffold(tf.keras.layers.Layer):
   instantiate the class with the config, or pass a class instance to
   `attention_cls`/`feedforward_cls`.
 
-  Arguments:
+  Args:
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate layer.
     intermediate_activation: Activation for the intermediate layer.

official/nlp/modeling/layers/transformer_xl.py

@@ -25,7 +25,7 @@ from official.nlp.modeling.layers import relative_attention
 def _cache_memory(current_state, previous_state, memory_length, reuse_length=0):
   """Caches hidden states into memory.
 
-  Arguments:
+  Args:
     current_state: `Tensor`, the current state.
     previous_state: `Tensor`, the previous state.
     memory_length: `int`, the number of tokens to cache.
@@ -228,7 +228,7 @@ class TransformerXLBlock(tf.keras.layers.Layer):
            target_mapping=None):
     """Implements `call` for the Layer.
 
-    Arguments:
+    Args:
       content_stream: `Tensor`, the input content stream. This is the standard
         input to Transformer XL and is commonly referred to as `h` in XLNet.
       content_attention_bias: Bias `Tensor` for content based attention of shape
@@ -476,7 +476,7 @@ class TransformerXL(tf.keras.layers.Layer):
            target_mapping=None):
     """Implements call() for the layer.
 
-    Arguments:
+    Args:
       content_stream: `Tensor`, the input content stream. This is the standard
         input to Transformer XL and is commonly referred to as `h` in XLNet.
       relative_position_encoding: Relative positional encoding `Tensor` of shape

official/nlp/modeling/models/bert_classifier.py

@@ -36,7 +36,7 @@ class BertClassifier(tf.keras.Model):
   *Note* that the model is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     network: A transformer network. This network should output a sequence output
       and a classification output. Furthermore, it should expose its embedding
       table via a "get_embedding_table" method.

official/nlp/modeling/models/bert_pretrainer.py

@@ -39,7 +39,7 @@ class BertPretrainer(tf.keras.Model):
   *Note* that the model is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     network: A transformer network. This network should output a sequence output
       and a classification output.
     num_classes: Number of classes to predict from the classification network.
@@ -165,7 +165,7 @@ class BertPretrainerV2(tf.keras.Model):
   Adds the masked language model head and optional classification heads upon the
   transformer encoder.
 
-  Arguments:
+  Args:
     encoder_network: A transformer network. This network should output a
       sequence output and a classification output.
     mlm_activation: The activation (if any) to use in the masked LM network. If

official/nlp/modeling/models/bert_span_labeler.py

@@ -34,7 +34,7 @@ class BertSpanLabeler(tf.keras.Model):
   *Note* that the model is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     network: A transformer network. This network should output a sequence output
       and a classification output. Furthermore, it should expose its embedding
       table via a "get_embedding_table" method.

official/nlp/modeling/models/bert_token_classifier.py

@@ -33,7 +33,7 @@ class BertTokenClassifier(tf.keras.Model):
   *Note* that the model is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     network: A transformer network. This network should output a sequence output
       and a classification output. Furthermore, it should expose its embedding
       table via a "get_embedding_table" method.

official/nlp/modeling/models/dual_encoder.py

@@ -31,7 +31,7 @@ class DualEncoder(tf.keras.Model):
   The DualEncoder allows a user to pass in a transformer stack, and build a dual
   encoder model based on the transformer stack.
 
-  Arguments:
+  Args:
     network: A transformer network which should output an encoding output.
     max_seq_length: The maximum allowed sequence length for transformer.
     normalize: If set to True, normalize the encoding produced by transfomer.

official/nlp/modeling/models/electra_pretrainer.py

@@ -39,7 +39,7 @@ class ElectraPretrainer(tf.keras.Model):
   *Note* that the model is constructed by Keras Subclass API, where layers are
   defined inside __init__ and call() implements the computation.
 
-  Arguments:
+  Args:
     generator_network: A transformer network for generator, this network should
       output a sequence output and an optional classification output.
     discriminator_network: A transformer network for discriminator, this network

official/nlp/modeling/models/seq2seq_transformer.py

@@ -57,7 +57,7 @@ class Seq2SeqTransformer(tf.keras.Model):
                **kwargs):
     """Initialize layers to build Transformer model.
 
-    Arguments:
+    Args:
       vocab_size: Size of vocabulary.
       embedding_width: Size of hidden layer for embedding.
       dropout_rate: Dropout probability.
@@ -359,7 +359,7 @@ class TransformerEncoder(tf.keras.layers.Layer):
     1. Self-attention layer
     2. Feedforward network (which is 2 fully-connected layers)
 
-  Arguments:
+  Args:
     num_layers: Number of layers.
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate (Feedforward) layer.
@@ -468,7 +468,7 @@ class TransformerDecoder(tf.keras.layers.Layer):
        the previous self-attention layer.
     3. Feedforward network (2 fully-connected layers)
 
-  Arguments:
+  Args:
     num_layers: Number of layers.
     num_attention_heads: Number of attention heads.
     intermediate_size: Size of the intermediate (Feedforward) layer.

official/nlp/modeling/models/xlnet.py

@@ -84,7 +84,7 @@ class XLNetPretrainer(tf.keras.Model):
   Transformer-XL encoder as described in "XLNet: Generalized Autoregressive
   Pretraining for Language Understanding" (https://arxiv.org/abs/1906.08237).
 
-  Arguments:
+  Args:
     network: An XLNet/Transformer-XL based network. This network should output a
       sequence output and list of `state` tensors.
     mlm_activation: The activation (if any) to use in the Masked LM network. If
@@ -163,7 +163,7 @@ class XLNetClassifier(tf.keras.Model):
   Note: This model does not use utilize the memory mechanism used in the
   original XLNet Classifier.
 
-  Arguments:
+  Args:
     network: An XLNet/Transformer-XL based network. This network should output a
       sequence output and list of `state` tensors.
     num_classes: Number of classes to predict from the classification network.
@@ -249,7 +249,7 @@ class XLNetSpanLabeler(tf.keras.Model):
   Transformer-XL encoder as described in "XLNet: Generalized Autoregressive
   Pretraining for Language Understanding" (https://arxiv.org/abs/1906.08237).
 
-  Arguments:
+  Args:
     network: A transformer network. This network should output a sequence output
       and a classification output. Furthermore, it should expose its embedding
       table via a "get_embedding_table" method.

official/nlp/modeling/networks/albert_encoder.py

@@ -39,7 +39,7 @@ class AlbertEncoder(tf.keras.Model):
 
   *Note* that the network is constructed by Keras Functional API.
 
-  Arguments:
+  Args:
     vocab_size: The size of the token vocabulary.
     embedding_width: The width of the word embeddings. If the embedding width is
       not equal to hidden size, embedding parameters will be factorized into two

official/nlp/modeling/networks/bert_encoder.py

@@ -41,7 +41,7 @@ class BertEncoder(keras_nlp.encoders.BertEncoder):
   *Note* that the network is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     vocab_size: The size of the token vocabulary.
     hidden_size: The size of the transformer hidden layers.
     num_layers: The number of transformer layers.

official/nlp/modeling/networks/classification.py

@@ -28,7 +28,7 @@ class Classification(tf.keras.Model):
   *Note* that the network is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     input_width: The innermost dimension of the input tensor to this network.
     num_classes: The number of classes that this network should classify to. If
       equal to 1, a regression problem is assumed.

official/nlp/modeling/networks/encoder_scaffold.py

@@ -49,7 +49,7 @@ class EncoderScaffold(tf.keras.Model):
   *Note* that the network is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     pooled_output_dim: The dimension of pooled output.
     pooler_layer_initializer: The initializer for the classification layer.
     embedding_cls: The class or instance to use to embed the input data. This

official/nlp/modeling/networks/mobile_bert_encoder.py

@@ -46,7 +46,7 @@ class MobileBERTEncoder(tf.keras.Model):
                **kwargs):
     """Class initialization.
 
-    Arguments:
+    Args:
       word_vocab_size: Number of words in the vocabulary.
       word_embed_size: Word embedding size.
       type_vocab_size: Number of word types.

official/nlp/modeling/networks/packed_sequence_embedding.py

@@ -33,7 +33,7 @@ class PackedSequenceEmbedding(tf.keras.Model):
   to (1) pack multiple sequences into one sequence and (2) allow additional
   "position_ids" as input.
 
-  Arguments:
+  Args:
     vocab_size: The size of the token vocabulary.
     type_vocab_size: The size of the type vocabulary.
     embedding_width: Width of token embeddings.
@@ -207,7 +207,7 @@ class PositionEmbeddingWithSubSeqMask(tf.keras.layers.Layer):
   can have a dynamic 1st dimension, while if `use_dynamic_slicing` is False the
   input size must be fixed.
 
-  Arguments:
+  Args:
     initializer: The initializer to use for the embedding weights. Defaults to
       "glorot_uniform".
     use_dynamic_slicing: Whether to use the dynamic slicing path.

official/nlp/modeling/networks/span_labeling.py

@@ -32,7 +32,7 @@ class SpanLabeling(tf.keras.Model):
   *Note* that the network is constructed by
   [Keras Functional API](https://keras.io/guides/functional_api/).
 
-  Arguments:
+  Args:
     input_width: The innermost dimension of the input tensor to this network.
     activation: The activation, if any, for the dense layer in this network.
     initializer: The initializer for the dense layer in this network. Defaults
@@ -123,7 +123,7 @@ class XLNetSpanLabeling(tf.keras.layers.Layer):
   **Note: `compute_with_beam_search` will not work with the Functional API
   (https://www.tensorflow.org/guide/keras/functional).
 
-  Arguments:
+  Args:
     input_width: The innermost dimension of the input tensor to this network.
     start_n_top: Beam size for span start.
     end_n_top: Beam size for span end.