Merge pull request #9090 from xinliupitt:master

PiperOrigin-RevId: 326550493

official/nlp/modeling/models/seq2seq_transformer.py

+# Copyright 2020 The TensorFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+"""Implement Seq2Seq Transformer model by TF official NLP library.
+
+Model paper: https://arxiv.org/pdf/1706.03762.pdf
+"""
+import math
+
+import tensorflow as tf
+from official.modeling import tf_utils
+from official.nlp.modeling import layers
+from official.nlp.modeling.ops import beam_search
+from official.nlp.transformer import metrics
+from official.nlp.transformer import model_utils
+
+EOS_ID = 1
+# pylint: disable=g-classes-have-attributes
+
+
+def create_model(params, is_train):
+  """Creates transformer model."""
+
+  encdec_kwargs = dict(
+      num_layers=params["num_hidden_layers"],
+      num_attention_heads=params["num_heads"],
+      intermediate_size=params["filter_size"],
+      activation="relu",
+      dropout_rate=params["relu_dropout"],
+      attention_dropout_rate=params["attention_dropout"],
+      use_bias=False,
+      norm_first=True,
+      norm_epsilon=1e-6,
+      intermediate_dropout=params["relu_dropout"])
+  encoder_layer = TransformerEncoder(**encdec_kwargs)
+  decoder_layer = TransformerDecoder(**encdec_kwargs)
+
+  model_kwargs = dict(
+      vocab_size=params["vocab_size"],
+      hidden_size=params["hidden_size"],
+      dropout_rate=params["layer_postprocess_dropout"],
+      padded_decode=params["padded_decode"],
+      num_replicas=params["num_replicas"],
+      decode_batch_size=params["decode_batch_size"],
+      decode_max_length=params["decode_max_length"],
+      dtype=params["dtype"],
+      extra_decode_length=params["extra_decode_length"],
+      num_heads=params["num_heads"],
+      num_layers=params["num_hidden_layers"],
+      beam_size=params["beam_size"],
+      alpha=params["alpha"],
+      encoder_layer=encoder_layer,
+      decoder_layer=decoder_layer,
+      name="transformer_v2")
+
+  with tf.name_scope("model"):
+    if is_train:
+      inputs = tf.keras.layers.Input((None,), dtype="int64", name="inputs")
+      targets = tf.keras.layers.Input((None,), dtype="int64", name="targets")
+      internal_model = Seq2SeqTransformer(**model_kwargs)
+      logits = internal_model([inputs, targets], training=is_train)
+      vocab_size = params["vocab_size"]
+      label_smoothing = params["label_smoothing"]
+      if params["enable_metrics_in_training"]:
+        logits = metrics.MetricLayer(vocab_size)([logits, targets])
+      logits = tf.keras.layers.Lambda(
+          lambda x: x, name="logits", dtype=tf.float32)(
+              logits)
+      model = tf.keras.Model([inputs, targets], logits)
+      loss = metrics.transformer_loss(logits, targets, label_smoothing,
+                                      vocab_size)
+      model.add_loss(loss)
+      return model
+
+    else:
+      inputs = tf.keras.layers.Input((None,), dtype="int64", name="inputs")
+      internal_model = Seq2SeqTransformer(**model_kwargs)
+      ret = internal_model([inputs], training=is_train)
+      outputs, scores = ret["outputs"], ret["scores"]
+      return tf.keras.Model(inputs, [outputs, scores])
+
+
+@tf.keras.utils.register_keras_serializable(package="Text")
+class Seq2SeqTransformer(tf.keras.Model):
+  """Transformer model with Keras.
+
+  Implemented as described in: https://arxiv.org/pdf/1706.03762.pdf
+
+  The Transformer model consists of an encoder and decoder. The input is an int
+  sequence (or a batch of sequences). The encoder produces a continuous
+  representation, and the decoder uses the encoder output to generate
+  probabilities for the output sequence.
+  """
+
+  def __init__(self,
+               vocab_size=33708,
+               hidden_size=512,
+               dropout_rate=0.0,
+               padded_decode=False,
+               num_replicas=1,
+               decode_batch_size=2048,
+               decode_max_length=97,
+               dtype=tf.float32,
+               extra_decode_length=0,
+               num_heads=8,
+               num_layers=6,
+               beam_size=4,
+               alpha=0.6,
+               encoder_layer=None,
+               decoder_layer=None,
+               name=None,
+               **kwargs):
+    """Initialize layers to build Transformer model.
+
+    Arguments:
+      vocab_size: Size of vocabulary.
+      hidden_size: Size of hidden layer for embedding.
+      dropout_rate: Dropout probability.
+      padded_decode: Whether to max_sequence_length padding is used. If set
+        False, max_sequence_length padding is not used.
+      num_replicas: Number of replicas for distribution strategy.
+      decode_batch_size: batch_size for decoding.
+      decode_max_length: maximum number of steps to decode a sequence.
+      dtype: data type.
+      extra_decode_length: Beam search will run extra steps to decode.
+      num_heads: Number of attention heads.
+      num_layers: Number of identical layers for Transformer architecture.
+      beam_size: Number of beams for beam search
+      alpha: The strength of length normalization for beam search.
+      encoder_layer: An initialized encoder layer.
+      decoder_layer: An initialized decoder layer.
+      name: name of the model.
+      **kwargs: other keyword arguments.
+    """
+    super(Seq2SeqTransformer, self).__init__(**kwargs)
+    self._vocab_size = vocab_size
+    self._hidden_size = hidden_size
+    self._dropout_rate = dropout_rate
+    self._padded_decode = padded_decode
+    self._num_replicas = num_replicas
+    self._decode_batch_size = decode_batch_size
+    self._decode_max_length = decode_max_length
+    self._dtype = dtype
+    self._extra_decode_length = extra_decode_length
+    self._num_heads = num_heads
+    self._num_layers = num_layers
+    self._beam_size = beam_size
+    self._alpha = alpha
+    self.embedding_lookup = layers.OnDeviceEmbedding(
+        vocab_size=self._vocab_size,
+        embedding_width=self._hidden_size,
+        initializer=tf.random_normal_initializer(
+            mean=0., stddev=self._hidden_size**-0.5),
+        use_scale=True)
+    self.encoder_layer = encoder_layer
+    self.decoder_layer = decoder_layer
+    self.position_embedding = layers.RelativePositionEmbedding(
+        hidden_size=self._hidden_size)
+    self.encoder_dropout = tf.keras.layers.Dropout(rate=self._dropout_rate)
+    self.decoder_dropout = tf.keras.layers.Dropout(rate=self._dropout_rate)
+
+  def get_config(self):
+    config = {
+        "vocab_size": self._vocab_size,
+        "hidden_size": self._hidden_size,
+        "dropout_rate": self._dropout_rate,
+        "padded_decode": self._padded_decode,
+        "num_replicas": self._num_replicas,
+        "decode_batch_size": self._decode_batch_size,
+        "decode_max_length": self._decode_max_length,
+        "dtype": self._dtype,
+        "extra_decode_length": self._extra_decode_length,
+        "num_heads": self._num_heads,
+        "num_layers": self._num_layers,
+        "beam_size": self._beam_size,
+        "alpha": self._alpha,
+        "encoder_layer": self.encoder_layer,
+        "decoder_layer": self.decoder_layer
+    }
+    base_config = super(Seq2SeqTransformer, self).get_config()
+    return dict(list(base_config.items()) + list(config.items()))
+
+  def call(self, inputs):
+    """Calculate target logits or inferred target sequences.
+
+    Args:
+      inputs: input tensor list of size 1 or 2.
+        First item, inputs: int tensor with shape [batch_size, input_length].
+        Second item (optional), targets: None or int tensor with shape
+          [batch_size, target_length].
+
+    Returns:
+      If targets is defined, then return logits for each word in the target
+      sequence. float tensor with shape [batch_size, target_length, vocab_size]
+      If target is none, then generate output sequence one token at a time.
+        returns a dictionary {
+          outputs: [batch_size, decoded length]
+          scores: [batch_size, float]}
+      Even when float16 is used, the output tensor(s) are always float32.
+
+    Raises:
+      NotImplementedError: If try to use padded decode method on CPU/GPUs.
+    """
+    if len(inputs) == 2:
+      inputs, targets = inputs[0], inputs[1]
+    else:
+      # Decoding path.
+      inputs, targets = inputs[0], None
+
+      # TODO(hongkuny): The check is not necessary. Fix this part.
+      if self._padded_decode:
+        if not self._num_replicas:
+          raise NotImplementedError(
+              "Padded decoding on CPU/GPUs is not supported.")
+        decode_batch_size = int(self._decode_batch_size / self._num_replicas)
+        inputs.set_shape([decode_batch_size, self._decode_max_length])
+
+    with tf.name_scope("Transformer"):
+      attention_bias = model_utils.get_padding_bias(inputs)
+      attention_bias = tf.cast(attention_bias, self._dtype)
+      with tf.name_scope("encode"):
+        # Prepare inputs to the layer stack by adding positional encodings and
+        # applying dropout.
+        embedded_inputs = self.embedding_lookup(inputs)
+        embedding_mask = tf.cast(
+            tf.not_equal(inputs, 0), self.embedding_lookup.embeddings.dtype)
+        embedded_inputs *= tf.expand_dims(embedding_mask, -1)
+        embedded_inputs = tf.cast(embedded_inputs, self._dtype)
+
+        # Attention_mask generation.
+        input_shape = tf_utils.get_shape_list(inputs, expected_rank=2)
+        attention_mask = tf.cast(
+            tf.reshape(
+                tf.not_equal(inputs, 0), [input_shape[0], 1, input_shape[1]]),
+            dtype=inputs.dtype)
+        broadcast_ones = tf.ones(
+            shape=[input_shape[0], input_shape[1], 1], dtype=inputs.dtype)
+        attention_mask = broadcast_ones * attention_mask
+
+        with tf.name_scope("add_pos_encoding"):
+          pos_encoding = self.position_embedding(inputs=embedded_inputs)
+          pos_encoding = tf.cast(pos_encoding, self._dtype)
+          encoder_inputs = embedded_inputs + pos_encoding
+
+        encoder_inputs = self.encoder_dropout(encoder_inputs)
+
+        encoder_outputs = self.encoder_layer(
+            encoder_inputs, attention_mask=attention_mask)
+
+      if targets is None:
+        encoder_decoder_attention_bias = attention_bias
+        encoder_outputs = tf.cast(encoder_outputs, self._dtype)
+        if self._padded_decode:
+          batch_size = encoder_outputs.shape.as_list()[0]
+          input_length = encoder_outputs.shape.as_list()[1]
+        else:
+          batch_size = tf.shape(encoder_outputs)[0]
+          input_length = tf.shape(encoder_outputs)[1]
+        max_decode_length = input_length + self._extra_decode_length
+        encoder_decoder_attention_bias = tf.cast(encoder_decoder_attention_bias,
+                                                 self._dtype)
+
+        symbols_to_logits_fn = self._get_symbols_to_logits_fn(max_decode_length)
+
+        # Create initial set of IDs that will be passed to symbols_to_logits_fn.
+        initial_ids = tf.zeros([batch_size], dtype=tf.int32)
+
+        # Create cache storing decoder attention values for each layer.
+        # pylint: disable=g-complex-comprehension
+        init_decode_length = (max_decode_length if self._padded_decode else 0)
+        num_heads = self._num_heads
+        dim_per_head = self._hidden_size // num_heads
+
+        cache = {
+            str(layer): {
+                "key":
+                    tf.zeros([
+                        batch_size, init_decode_length, num_heads, dim_per_head
+                    ],
+                             dtype=self._dtype),
+                "value":
+                    tf.zeros([
+                        batch_size, init_decode_length, num_heads, dim_per_head
+                    ],
+                             dtype=self._dtype)
+            } for layer in range(self._num_layers)
+        }
+
+        # pylint: enable=g-complex-comprehension
+
+        # Add encoder output and attention bias to the cache.
+        cache["encoder_outputs"] = encoder_outputs
+        cache["encoder_decoder_attention_bias"] = encoder_decoder_attention_bias
+
+        # Use beam search to find the top beam_size sequences and scores.
+        decoded_ids, scores = beam_search.sequence_beam_search(
+            symbols_to_logits_fn=symbols_to_logits_fn,
+            initial_ids=initial_ids,
+            initial_cache=cache,
+            vocab_size=self._vocab_size,
+            beam_size=self._beam_size,
+            alpha=self._alpha,
+            max_decode_length=max_decode_length,
+            eos_id=EOS_ID,
+            padded_decode=self._padded_decode,
+            dtype=self._dtype)
+
+        # Get the top sequence for each batch element
+        top_decoded_ids = decoded_ids[:, 0, 1:]
+        top_scores = scores[:, 0]
+
+        return {"outputs": top_decoded_ids, "scores": top_scores}
+
+      else:
+        with tf.name_scope("decode"):
+          decoder_inputs = self.embedding_lookup(targets)
+          embedding_mask = tf.cast(
+              tf.not_equal(targets, 0), self.embedding_lookup.embeddings.dtype)
+          decoder_inputs *= tf.expand_dims(embedding_mask, -1)
+          decoder_inputs = tf.cast(decoder_inputs, self._dtype)
+          with tf.name_scope("shift_targets"):
+            # Shift targets to the right, and remove the last element
+            decoder_inputs = tf.pad(decoder_inputs,
+                                    [[0, 0], [1, 0], [0, 0]])[:, :-1, :]
+          with tf.name_scope("add_pos_encoding"):
+            length = tf.shape(decoder_inputs)[1]
+            pos_encoding = self.position_embedding(decoder_inputs)
+            pos_encoding = tf.cast(pos_encoding, self._dtype)
+            decoder_inputs += pos_encoding
+
+          decoder_inputs = self.decoder_dropout(decoder_inputs)
+
+          decoder_shape = tf_utils.get_shape_list(
+              decoder_inputs, expected_rank=3)
+          batch_size = decoder_shape[0]
+          decoder_length = decoder_shape[1]
+
+          self_attention_mask = tf.linalg.band_part(
+              tf.ones([length, length], dtype=tf.float32), -1, 0)
+          self_attention_mask = tf.reshape(self_attention_mask,
+                                           [1, length, length])
+          self_attention_mask = tf.tile(self_attention_mask, [batch_size, 1, 1])
+
+          attention_mask = tf.cast(
+              tf.expand_dims(tf.not_equal(inputs, 0), axis=1),
+              dtype=inputs.dtype)
+          attention_mask = tf.tile(attention_mask, [1, decoder_length, 1])
+
+          outputs = self.decoder_layer(
+              decoder_inputs,
+              encoder_outputs,
+              memory_mask=self_attention_mask,
+              target_mask=attention_mask)
+          logits = embedding_linear(self.embedding_lookup.embeddings, outputs)
+          logits = tf.cast(logits, tf.float32)
+
+        return logits
+
+  def _get_symbols_to_logits_fn(self, max_decode_length):
+    """Returns a decoding function that calculates logits of the next tokens."""
+    timing_signal = self.position_embedding(
+        inputs=None, length=max_decode_length + 1)
+    decoder_self_attention_bias = model_utils.get_decoder_self_attention_bias(
+        max_decode_length, dtype=self._dtype)
+
+    def symbols_to_logits_fn(ids, i, cache):
+      """Generate logits for next potential IDs.
+
+      Args:
+        ids: Current decoded sequences. int tensor with shape [batch_size *
+          beam_size, i + 1].
+        i: Loop index.
+        cache: dictionary of values storing the encoder output, encoder-decoder
+          attention bias, and previous decoder attention values.
+
+      Returns:
+        Tuple of
+          (logits with shape [batch_size * beam_size, vocab_size],
+           updated cache values)
+      """
+      # Set decoder input to the last generated IDs
+      decoder_input = ids[:, -1:]
+
+      # Preprocess decoder input by getting embeddings and adding timing signal.
+      # decoder_input = self.embedding_softmax_layer(decoder_input)
+      source_decoder_input = decoder_input
+      decoder_input = self.embedding_lookup(decoder_input)
+      embedding_mask = tf.cast(
+          tf.not_equal(source_decoder_input, 0),
+          self.embedding_lookup.embeddings.dtype)
+      decoder_input *= tf.expand_dims(embedding_mask, -1)
+
+      if self._padded_decode:
+        timing_signal_shape = timing_signal.shape.as_list()
+        decoder_input += tf.slice(timing_signal, [i, 0],
+                                  [1, timing_signal_shape[1]])
+
+        bias_shape = decoder_self_attention_bias.shape.as_list()
+        self_attention_bias = tf.slice(
+            decoder_self_attention_bias, [0, 0, i, 0],
+            [bias_shape[0], bias_shape[1], 1, bias_shape[3]])
+      else:
+        decoder_input += timing_signal[i:i + 1]
+
+        self_attention_bias = decoder_self_attention_bias[:, :, i:i + 1, :i + 1]
+
+      decoder_shape = tf_utils.get_shape_list(decoder_input, expected_rank=3)
+      batch_size = decoder_shape[0]
+      decoder_length = decoder_shape[1]
+
+      attention_bias = cache.get("encoder_decoder_attention_bias")
+      attention_bias = tf.where(attention_bias < 0,
+                                tf.zeros_like(attention_bias),
+                                tf.ones_like(attention_bias))
+      attention_bias = tf.squeeze(attention_bias, axis=[1])
+      attention_mask = tf.tile(attention_bias, [1, decoder_length, 1])
+
+      self_attention_bias = tf.where(self_attention_bias < 0,
+                                     tf.zeros_like(self_attention_bias),
+                                     tf.ones_like(self_attention_bias))
+      self_attention_bias = tf.squeeze(self_attention_bias, axis=[1])
+      self_attention_mask = tf.tile(self_attention_bias, [batch_size, 1, 1])
+
+      decoder_outputs = self.decoder_layer(
+          decoder_input,
+          cache.get("encoder_outputs"),
+          memory_mask=self_attention_mask,
+          target_mask=attention_mask,
+          cache=cache,
+          decode_loop_step=i if self._padded_decode else None)
+
+      logits = embedding_linear(self.embedding_lookup.embeddings,
+                                decoder_outputs)
+      logits = tf.squeeze(logits, axis=[1])
+      return logits, cache
+
+    return symbols_to_logits_fn
+
+
+class TransformerEncoder(tf.keras.layers.Layer):
+  """Transformer encoder.
+
+  Transformer encoder is made up of N identical layers. Each layer is composed
+  of the sublayers:
+    1. Self-attention layer
+    2. Feedforward network (which is 2 fully-connected layers)
+
+  Arguments:
+    num_layers: Number of layers.
+    num_attention_heads: Number of attention heads.
+    intermediate_size: Size of the intermediate (Feedforward) layer.
+    activation: Activation for the intermediate layer.
+    dropout_rate: Dropout probability.
+    attention_dropout_rate: Dropout probability for attention layers.
+    use_bias: Whether to enable use_bias in attention layer. If set False,
+      use_bias in attention layer is disabled.
+    norm_first: Whether to normalize inputs to attention and intermediate dense
+      layers. If set False, output of attention and intermediate dense layers is
+      normalized.
+    norm_epsilon: Epsilon value to initialize normalization layers.
+    intermediate_dropout: Dropout probability for intermediate_dropout_layer.
+  """
+
+  def __init__(self,
+               num_layers=6,
+               num_attention_heads=8,
+               intermediate_size=2048,
+               activation="relu",
+               dropout_rate=0.0,
+               attention_dropout_rate=0.0,
+               use_bias=False,
+               norm_first=True,
+               norm_epsilon=1e-6,
+               intermediate_dropout=0.0,
+               **kwargs):
+    super(TransformerEncoder, self).__init__(**kwargs)
+    self._num_layers = num_layers
+    self._num_attention_heads = num_attention_heads
+    self._intermediate_size = intermediate_size
+    self._activation = activation
+    self._dropout_rate = dropout_rate
+    self._attention_dropout_rate = attention_dropout_rate
+    self._use_bias = use_bias
+    self._norm_first = norm_first
+    self._norm_epsilon = norm_epsilon
+    self._intermediate_dropout = intermediate_dropout
+
+  def build(self, input_shape):
+    """Implements build() for the layer."""
+    self.encoder_layers = []
+    for i in range(self._num_layers):
+      self.encoder_layers.append(
+          layers.Transformer(
+              num_attention_heads=self._num_attention_heads,
+              intermediate_size=self._intermediate_size,
+              intermediate_activation=self._activation,
+              dropout_rate=self._dropout_rate,
+              attention_dropout_rate=self._attention_dropout_rate,
+              use_bias=self._use_bias,
+              norm_first=self._norm_first,
+              norm_epsilon=self._norm_epsilon,
+              intermediate_dropout=self._intermediate_dropout,
+              attention_initializer=attention_initializer(input_shape[2]),
+              name=("layer_%d" % i)))
+    self.output_normalization = tf.keras.layers.LayerNormalization(
+        epsilon=self._norm_epsilon, dtype="float32")
+    super(TransformerEncoder, self).build(input_shape)
+
+  def get_config(self):
+    config = {
+        "num_layers": self._num_layers,
+        "num_attention_heads": self._num_attention_heads,
+        "intermediate_size": self._intermediate_size,
+        "activation": self._activation,
+        "dropout_rate": self._dropout_rate,
+        "attention_dropout_rate": self._attention_dropout_rate,
+        "use_bias": self._use_bias,
+        "norm_first": self._norm_first,
+        "norm_epsilon": self._norm_epsilon,
+        "intermediate_dropout": self._intermediate_dropout
+    }
+    base_config = super(TransformerEncoder, self).get_config()
+    return dict(list(base_config.items()) + list(config.items()))
+
+  def call(self, encoder_inputs, attention_mask=None):
+    """Return the output of the encoder.
+
+    Args:
+      encoder_inputs: tensor with shape [batch_size, input_length, hidden_size]
+      attention_mask: mask for the encoder self-attention layer. [batch_size,
+        input_length, input_length]
+
+    Returns:
+      Output of encoder.
+      float32 tensor with shape [batch_size, input_length, hidden_size]
+    """
+    for layer_idx in range(self._num_layers):
+      encoder_inputs = self.encoder_layers[layer_idx](
+          [encoder_inputs, attention_mask])
+
+    output_tensor = encoder_inputs
+    output_tensor = self.output_normalization(output_tensor)
+
+    return output_tensor
+
+
+class TransformerDecoder(tf.keras.layers.Layer):
+  """Transformer decoder.
+
+  Like the encoder, the decoder is made up of N identical layers.
+  Each layer is composed of the sublayers:
+    1. Self-attention layer
+    2. Multi-headed attention layer combining encoder outputs with results from
+       the previous self-attention layer.
+    3. Feedforward network (2 fully-connected layers)
+
+  Arguments:
+    num_layers: Number of layers.
+    num_attention_heads: Number of attention heads.
+    intermediate_size: Size of the intermediate (Feedforward) layer.
+    activation: Activation for the intermediate layer.
+    dropout_rate: Dropout probability.
+    attention_dropout_rate: Dropout probability for attention layers.
+    use_bias: Whether to enable use_bias in attention layer. If set False,
+      use_bias in attention layer is disabled.
+    norm_first: Whether to normalize inputs to attention and intermediate dense
+      layers. If set False, output of attention and intermediate dense layers is
+      normalized.
+    norm_epsilon: Epsilon value to initialize normalization layers.
+    intermediate_dropout: Dropout probability for intermediate_dropout_layer.
+  """
+
+  def __init__(self,
+               num_layers=6,
+               num_attention_heads=8,
+               intermediate_size=2048,
+               activation="relu",
+               dropout_rate=0.0,
+               attention_dropout_rate=0.0,
+               use_bias=False,
+               norm_first=True,
+               norm_epsilon=1e-6,
+               intermediate_dropout=0.0,
+               **kwargs):
+    super(TransformerDecoder, self).__init__(**kwargs)
+    self._num_layers = num_layers
+    self._num_attention_heads = num_attention_heads
+    self._intermediate_size = intermediate_size
+    self._activation = activation
+    self._dropout_rate = dropout_rate
+    self._attention_dropout_rate = attention_dropout_rate
+    self._use_bias = use_bias
+    self._norm_first = norm_first
+    self._norm_epsilon = norm_epsilon
+    self._intermediate_dropout = intermediate_dropout
+
+  def build(self, input_shape):
+    """Implements build() for the layer."""
+    self.decoder_layers = []
+    for i in range(self._num_layers):
+      self.decoder_layers.append(
+          layers.TransformerDecoderLayer(
+              num_attention_heads=self._num_attention_heads,
+              intermediate_size=self._intermediate_size,
+              intermediate_activation=self._activation,
+              dropout_rate=self._dropout_rate,
+              attention_dropout_rate=self._attention_dropout_rate,
+              use_bias=self._use_bias,
+              norm_first=self._norm_first,
+              norm_epsilon=self._norm_epsilon,
+              intermediate_dropout=self._intermediate_dropout,
+              attention_initializer=attention_initializer(input_shape[2]),
+              name=("layer_%d" % i)))
+    self.output_normalization = tf.keras.layers.LayerNormalization(
+        epsilon=1e-6, dtype="float32")
+    super(TransformerDecoder, self).build(input_shape)
+
+  def get_config(self):
+    config = {
+        "num_layers": self._num_layers,
+        "num_attention_heads": self._num_attention_heads,
+        "intermediate_size": self._intermediate_size,
+        "activation": self._activation,
+        "dropout_rate": self._dropout_rate,
+        "attention_dropout_rate": self._attention_dropout_rate,
+        "use_bias": self._use_bias,
+        "norm_first": self._norm_first,
+        "norm_epsilon": self._norm_epsilon,
+        "intermediate_dropout": self._intermediate_dropout
+    }
+    base_config = super(TransformerDecoder, self).get_config()
+    return dict(list(base_config.items()) + list(config.items()))
+
+  def call(self,
+           target,
+           memory,
+           memory_mask=None,
+           target_mask=None,
+           cache=None,
+           decode_loop_step=None):
+    """Return the output of the decoder layer stacks.
+
+    Args:
+      target: A tensor with shape [batch_size, target_length, hidden_size].
+      memory: A tensor with shape [batch_size, input_length, hidden_size]
+      memory_mask: A tensor with shape [batch_size, target_len, target_length],
+        the mask for decoder self-attention layer.
+      target_mask: A tensor with shape [batch_size, target_length, input_length]
+        which is the mask for encoder-decoder attention layer.
+      cache: (Used for fast decoding) A nested dictionary storing previous
+        decoder self-attention values. The items are:
+          {layer_n: {"k": A tensor with shape [batch_size, i, key_channels],
+                     "v": A tensor with shape [batch_size, i, value_channels]},
+                       ...}
+      decode_loop_step: An integer, the step number of the decoding loop. Used
+        only for autoregressive inference on TPU.
+
+    Returns:
+      Output of decoder.
+      float32 tensor with shape [batch_size, target_length, hidden_size]
+    """
+
+    output_tensor = target
+    for layer_idx in range(self._num_layers):
+      transformer_inputs = [output_tensor, memory, target_mask, memory_mask]
+      # Gets the cache for decoding.
+      if cache is None:
+        output_tensor, _ = self.decoder_layers[layer_idx](transformer_inputs)
+      else:
+        cache_layer_idx = str(layer_idx)
+        output_tensor, cache[cache_layer_idx] = self.decoder_layers[layer_idx](
+            transformer_inputs,
+            cache=cache[cache_layer_idx],
+            decode_loop_step=decode_loop_step)
+    return self.output_normalization(output_tensor)
+
+
+def embedding_linear(embedding_matrix, x):
+  """Uses embeddings as linear transformation weights."""
+  with tf.name_scope("presoftmax_linear"):
+    batch_size = tf.shape(x)[0]
+    length = tf.shape(x)[1]
+    hidden_size = tf.shape(x)[2]
+    vocab_size = tf.shape(embedding_matrix)[0]
+
+    x = tf.reshape(x, [-1, hidden_size])
+    logits = tf.matmul(x, embedding_matrix, transpose_b=True)
+
+    return tf.reshape(logits, [batch_size, length, vocab_size])
+
+
+def attention_initializer(hidden_size):
+  """Initializer for attention layers in Seq2SeqTransformer."""
+  limit = math.sqrt(6.0 / (hidden_size + hidden_size))
+  return tf.keras.initializers.RandomUniform(minval=-limit, maxval=limit)

official/nlp/modeling/models/seq2seq_transformer_test.py

+# Copyright 2020 The TensorFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+"""Test Transformer model."""
+
+import tensorflow as tf
+
+from official.nlp.modeling.models import seq2seq_transformer
+from official.nlp.transformer import model_params
+
+
+class Seq2SeqTransformerTest(tf.test.TestCase):
+
+  def setUp(self):
+    super().setUp()
+    self.params = params = model_params.TINY_PARAMS
+    params["batch_size"] = params["default_batch_size"] = 16
+    params["hidden_size"] = 12
+    params["num_hidden_layers"] = 2
+    params["filter_size"] = 14
+    params["num_heads"] = 2
+    params["vocab_size"] = 41
+    params["extra_decode_length"] = 2
+    params["beam_size"] = 3
+    params["dtype"] = tf.float32
+
+  def test_create_model_train(self):
+    model = seq2seq_transformer.create_model(self.params, True)
+    inputs, outputs = model.inputs, model.outputs
+    self.assertLen(inputs, 2)
+    self.assertLen(outputs, 1)
+    self.assertEqual(inputs[0].shape.as_list(), [None, None])
+    self.assertEqual(inputs[0].dtype, tf.int64)
+    self.assertEqual(inputs[1].shape.as_list(), [None, None])
+    self.assertEqual(inputs[1].dtype, tf.int64)
+    self.assertEqual(outputs[0].shape.as_list(), [None, None, 41])
+    self.assertEqual(outputs[0].dtype, tf.float32)
+
+  def test_create_model_not_train(self):
+    model = seq2seq_transformer.create_model(self.params, False)
+    inputs, outputs = model.inputs, model.outputs
+    self.assertEqual(len(inputs), 1)
+    self.assertEqual(len(outputs), 2)
+    self.assertEqual(inputs[0].shape.as_list(), [None, None])
+    self.assertEqual(inputs[0].dtype, tf.int64)
+    self.assertEqual(outputs[0].shape.as_list(), [None, None])
+    self.assertEqual(outputs[0].dtype, tf.int32)
+    self.assertEqual(outputs[1].shape.as_list(), [None])
+    self.assertEqual(outputs[1].dtype, tf.float32)
+
+
+if __name__ == "__main__":
+  tf.test.main()