def get_embedding_output(self):
"""Gets output of the embedding lookup (i.e., input to the transformer).
Returns:
float Tensor of shape [batch_size, seq_length, hidden_size] corresponding
to the output of the embedding layer, after summing the word
embeddings with the positional embeddings and the token type embeddings,
then performing layer normalization. This is the input to the transformer.
"""
return bf.r_cast(self.embedding_output)
def create_model_top(bert_config, is_training, input_ids, input_mask,
segment_ids, label_ids, num_labels,
use_one_hot_embeddings):
if bert_config.precision == "bfloat16":
with tf.contrib.tpu.bfloat16_scope():
(total_loss, per_example_loss, logits,
probabilities) = create_model(bert_config, is_training, input_ids,
input_mask, segment_ids, label_ids,
num_labels, use_one_hot_embeddings)
logits = bf.r_cast(logits)
return (total_loss, per_example_loss, logits, probabilities)
else:
(total_loss, per_example_loss, logits,
probabilities) = create_model(bert_config, is_training, input_ids,
input_mask, segment_ids, label_ids,
num_labels, use_one_hot_embeddings)
return (total_loss, per_example_loss, logits, probabilities)
def create_model_top(bert_config,
is_training,
input_ids,
input_mask,
segment_ids,
label_ids,
num_labels,
use_one_hot_embeddings,
frozen_graph=None):
if frozen_graph and not is_training:
inputs = {
'input_ids': input_ids,
'input_mask': input_mask,
'segment_ids': segment_ids,
'label_ids': label_ids
}
outputs = ['loss/Mean', 'loss/Sum', 'loss/BiasAdd', 'loss/Softmax']
outputs = ['{}:0'.format(name) for name in outputs]
with tf.io.gfile.GFile(frozen_graph, "rb") as f:
graph_def = tf.compat.v1.GraphDef()
graph_def.ParseFromString(f.read())
return tf.graph_util.import_graph_def(graph_def,
inputs,
outputs,
name="")
if bert_config.precision == "bfloat16":
with tf.compat.v1.tpu.bfloat16_scope():
(total_loss, per_example_loss, logits,
probabilities) = create_model(bert_config, is_training, input_ids,
input_mask, segment_ids, label_ids,
num_labels, use_one_hot_embeddings)
logits = bf.r_cast(logits)
return (total_loss, per_example_loss, logits, probabilities)
else:
(total_loss, per_example_loss, logits,
probabilities) = create_model(bert_config, is_training, input_ids,
input_mask, segment_ids, label_ids,
num_labels, use_one_hot_embeddings)
return (total_loss, per_example_loss, logits, probabilities)
def attention_layer(from_tensor,
to_tensor,
attention_mask=None,
num_attention_heads=1,
size_per_head=512,
query_act=None,
key_act=None,
value_act=None,
attention_probs_dropout_prob=0.0,
initializer_range=0.02,
do_return_2d_tensor=False,
batch_size=None,
from_seq_length=None,
to_seq_length=None):
"""Performs multi-headed attention from `from_tensor` to `to_tensor`.
This is an implementation of multi-headed attention based on "Attention
is all you Need". If `from_tensor` and `to_tensor` are the same, then
this is self-attention. Each timestep in `from_tensor` attends to the
corresponding sequence in `to_tensor`, and returns a fixed-with vector.
This function first projects `from_tensor` into a "query" tensor and
`to_tensor` into "key" and "value" tensors. These are (effectively) a list
of tensors of length `num_attention_heads`, where each tensor is of shape
[batch_size, seq_length, size_per_head].
Then, the query and key tensors are dot-producted and scaled. These are
softmaxed to obtain attention probablpoties. The value tensors are then
interpolated by these probablpoties, then concatenated back to a single
tensor and returned.
In practice, the multi-headed attention are done with transposes and
reshapes rather than actual separate tensors.
Args:
from_tensor: float Tensor of shape [batch_size, from_seq_length,
from_width].
to_tensor: float Tensor of shape [batch_size, to_seq_length, to_width].
attention_mask: (optional) int32 Tensor of shape [batch_size,
from_seq_length, to_seq_length]. The values should be 1 or 0. The
attention scores will effectively be set to -infinity for any positions in
the mask that are 0, and will be unchanged for positions that are 1.
num_attention_heads: int. Number of attention heads.
size_per_head: int. Size of each attention head.
query_act: (optional) Activation function for the query transform.
key_act: (optional) Activation function for the key transform.
value_act: (optional) Activation function for the value transform.
attention_probs_dropout_prob: (optional) float. Dropout probablpoty of the
attention probablpoties.
initializer_range: float. Range of the weight initializer.
do_return_2d_tensor: bool. If True, the output will be of shape [batch_size
* from_seq_length, num_attention_heads * size_per_head]. If False, the
output will be of shape [batch_size, from_seq_length, num_attention_heads
* size_per_head].
batch_size: (Optional) int. If the input is 2D, this might be the batch size
of the 3D version of the `from_tensor` and `to_tensor`.
from_seq_length: (Optional) If the input is 2D, this might be the seq length
of the 3D version of the `from_tensor`.
to_seq_length: (Optional) If the input is 2D, this might be the seq length
of the 3D version of the `to_tensor`.
Returns:
float Tensor of shape [batch_size, from_seq_length,
num_attention_heads * size_per_head]. (If `do_return_2d_tensor` is
true, this will be of shape [batch_size * from_seq_length,
num_attention_heads * size_per_head]).
Raises:
ValueError: Any of the arguments or tensor shapes are invalid.
"""
def transpose_for_scores(input_tensor, batch_size, num_attention_heads,
seq_length, width):
output_tensor = tf.reshape(
input_tensor, [batch_size, seq_length, num_attention_heads, width])
output_tensor = tf.transpose(a=output_tensor, perm=[0, 2, 1, 3])
return output_tensor
from_shape = get_shape_list(from_tensor, expected_rank=[2, 3])
to_shape = get_shape_list(to_tensor, expected_rank=[2, 3])
if len(from_shape) != len(to_shape):
raise ValueError(
"The rank of `from_tensor` must match the rank of `to_tensor`.")
if len(from_shape) == 3:
batch_size = from_shape[0]
from_seq_length = from_shape[1]
to_seq_length = to_shape[1]
elif len(from_shape) == 2:
if (batch_size is None or from_seq_length is None
or to_seq_length is None):
raise ValueError(
"When passing in rank 2 tensors to attention_layer, the values "
"for `batch_size`, `from_seq_length`, and `to_seq_length` "
"must all be specified.")
# Scalar dimensions referenced here:
# B = batch size (number of sequences)
# F = `from_tensor` sequence length
# T = `to_tensor` sequence length
# N = `num_attention_heads`
# H = `size_per_head`
from_tensor_2d = reshape_to_matrix(from_tensor)
to_tensor_2d = reshape_to_matrix(to_tensor)
# `query_layer` = [B*F, N*H]
query_layer = tf.compat.v1.layers.dense(
from_tensor_2d,
num_attention_heads * size_per_head,
activation=query_act,
name="query",
kernel_initializer=create_initializer(initializer_range))
# `key_layer` = [B*T, N*H]
key_layer = tf.compat.v1.layers.dense(
to_tensor_2d,
num_attention_heads * size_per_head,
activation=key_act,
name="key",
kernel_initializer=create_initializer(initializer_range))
# `value_layer` = [B*T, N*H]
value_layer = tf.compat.v1.layers.dense(
to_tensor_2d,
num_attention_heads * size_per_head,
activation=value_act,
name="value",
kernel_initializer=create_initializer(initializer_range))
# `query_layer` = [B, N, F, H]
query_layer = transpose_for_scores(query_layer, batch_size,
num_attention_heads, from_seq_length,
size_per_head)
# `key_layer` = [B, N, T, H]
key_layer = transpose_for_scores(key_layer, batch_size,
num_attention_heads, to_seq_length,
size_per_head)
# Take the dot product between "query" and "key" to get the raw
# attention scores.
# `attention_scores` = [B, N, F, T]
attention_scores = tf.matmul(query_layer, key_layer, transpose_b=True)
attention_scores = tf.multiply(
attention_scores, bf.r_cast(1.0 / math.sqrt(float(size_per_head))))
if attention_mask is not None:
# `attention_mask` = [B, 1, F, T]
attention_mask = tf.expand_dims(attention_mask, axis=[1])
# Since attention_mask is 1.0 for positions we want to attend and 0.0 for
# masked positions, this operation will create a tensor which is 0.0 for
# positions we want to attend and -10000.0 for masked positions.
adder = (1.0 - tf.cast(attention_mask, from_tensor.dtype)) * -10000.0
# Since we are adding it to the raw scores before the softmax, this is
# effectively the same as removing these entirely.
attention_scores += adder
# Normalize the attention scores to probablpoties.
# `attention_probs` = [B, N, F, T]
attention_probs = bf.softmax(attention_scores)
# This is actually dropping out entire tokens to attend to, which might
# seem a bit unusual, but is taken from the original Transformer paper.
attention_probs = dropout(attention_probs, attention_probs_dropout_prob)
# `value_layer` = [B, T, N, H]
value_layer = tf.reshape(
value_layer,
[batch_size, to_seq_length, num_attention_heads, size_per_head])
# `value_layer` = [B, N, T, H]
value_layer = tf.transpose(a=value_layer, perm=[0, 2, 1, 3])
# `context_layer` = [B, N, F, H]
context_layer = tf.matmul(attention_probs, value_layer)
# `context_layer` = [B, F, N, H]
context_layer = tf.transpose(a=context_layer, perm=[0, 2, 1, 3])
if do_return_2d_tensor:
# `context_layer` = [B*F, N*H]
context_layer = tf.reshape(context_layer, [
batch_size * from_seq_length, num_attention_heads * size_per_head
])
else:
# `context_layer` = [B, F, N*H]
context_layer = tf.reshape(
context_layer,
[batch_size, from_seq_length, num_attention_heads * size_per_head])
return context_layer
def get_embedding_table(self):
if self.bf16_scope == True:
return (self.embedding_table)
else:
return bf.r_cast(self.embedding_table)
def __init__(self,
config,
is_training,
input_ids,
input_mask=None,
token_type_ids=None,
use_one_hot_embeddings=False,
scope=None):
"""Constructor for BertModel.
Args:
config: `BertConfig` instance.
is_training: bool. true for training model, false for eval model. Controls
whether dropout will be applied.
input_ids: int32 Tensor of shape [batch_size, seq_length].
input_mask: (optional) int32 Tensor of shape [batch_size, seq_length].
token_type_ids: (optional) int32 Tensor of shape [batch_size, seq_length].
use_one_hot_embeddings: (optional) bool. Whether to use one-hot word
embeddings or tf.embedding_lookup() for the word embeddings.
scope: (optional) variable scope. Defaults to "bert".
Raises:
ValueError: The config is invalid or one of the input tensor shapes
is invalid.
"""
# Flags for BF16 CPU
self.bf16_scope = False
if config.precision == "bfloat16":
bf.set_global_precision(tf.bfloat16)
if config.new_bf16_scope:
self.bf16_scope = True
bf.set_global_flags(optimized_softmax=config.optimized_softmax,
experimental_gelu=config.experimental_gelu)
config = copy.deepcopy(config)
if not is_training:
config.hidden_dropout_prob = 0.0
config.attention_probs_dropout_prob = 0.0
input_shape = get_shape_list(input_ids, expected_rank=2)
batch_size = input_shape[0]
seq_length = input_shape[1]
if input_mask is None:
input_mask = tf.ones(shape=[batch_size, seq_length],
dtype=tf.int32)
if token_type_ids is None:
token_type_ids = tf.zeros(shape=[batch_size, seq_length],
dtype=tf.int32)
with tf.compat.v1.variable_scope(scope, default_name="bert"):
with tf.compat.v1.variable_scope("embeddings"):
# Perform embedding lookup on the word ids.
(self.embedding_output,
self.embedding_table) = embedding_lookup(
input_ids=input_ids,
vocab_size=config.vocab_size,
embedding_size=config.hidden_size,
initializer_range=config.initializer_range,
word_embedding_name="word_embeddings",
use_one_hot_embeddings=use_one_hot_embeddings)
# Add positional embeddings and token type embeddings, then layer
# normalize and perform dropout.
self.embedding_output = embedding_postprocessor(
input_tensor=self.embedding_output,
use_token_type=True,
token_type_ids=token_type_ids,
token_type_vocab_size=config.type_vocab_size,
token_type_embedding_name="token_type_embeddings",
use_position_embeddings=True,
position_embedding_name="position_embeddings",
initializer_range=config.initializer_range,
max_position_embeddings=config.max_position_embeddings,
dropout_prob=config.hidden_dropout_prob)
with tf.compat.v1.variable_scope("encoder"):
# This converts a 2D mask of shape [batch_size, seq_length] to a 3D
# mask of shape [batch_size, seq_length, seq_length] which is used
# for the attention scores.
attention_mask = create_attention_mask_from_input_mask(
input_ids, input_mask)
# Run the stacked transformer.
# `sequence_output` shape = [batch_size, seq_length, hidden_size].
# Cast is used to cover bfloat16
input_tensor = bf.r_cast(self.embedding_output)
self.all_encoder_layers = transformer_model(
input_tensor=input_tensor,
attention_mask=attention_mask,
hidden_size=config.hidden_size,
num_hidden_layers=config.num_hidden_layers,
num_attention_heads=config.num_attention_heads,
intermediate_size=config.intermediate_size,
intermediate_act_fn=get_activation(config.hidden_act),
hidden_dropout_prob=config.hidden_dropout_prob,
attention_probs_dropout_prob=config.
attention_probs_dropout_prob,
initializer_range=config.initializer_range,
do_return_all_layers=True)
self.sequence_output = self.all_encoder_layers[-1]
# The "pooler" converts the encoded sequence tensor of shape
# [batch_size, seq_length, hidden_size] to a tensor of shape
# [batch_size, hidden_size]. This is necessary for segment-level
# (or segment-pair-level) classification tasks where we need a fixed
# dimensional representation of the segment.
with tf.compat.v1.variable_scope("pooler"):
# We "pool" the model by simply taking the hidden state corresponding
# to the first token. We assume that this has been pre-trained
first_token_tensor = tf.squeeze(self.sequence_output[:,
0:1, :],
axis=1)
self.pooled_output = tf.compat.v1.layers.dense(
first_token_tensor,
config.hidden_size,
activation=bf.tanh,
kernel_initializer=create_initializer(
config.initializer_range))
def get_embedding_table(self):
return bf.r_cast(self.embedding_table)
