def get_ifm_logits(features, feature_columns, shared_feature_vectors, units,
is_training, extra_options):
with tf.variable_scope('ifm'):
_check_ifm_args(extra_options)
use_shared_embedding = extra_options['ifm_use_shared_embedding']
use_project = extra_options['ifm_use_project']
project_size = extra_options['ifm_project_size']
hidden_unit = extra_options['ifm_hidden_unit']
field_dim = extra_options['ifm_field_dim']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
y = _ifm(feature_vectors, hidden_unit, field_dim, reduce_sum=True)
with tf.variable_scope('logits') as logits_scope:
logits = y
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def _model_fn(features, labels, mode, config):
head = head_lib._binary_logistic_or_multi_class_head( # pylint: disable=protected-access
n_classes, weight_column, label_vocabulary, loss_reduction,
loss_fn)
is_training = (mode == tf.estimator.ModeKeys.TRAIN)
net = tf.feature_column.input_layer(features, feature_columns)
inputs = _attention_layer(features, attention_columns, is_training)
tf.logging.info('attention outputs = {}'.format(inputs))
inputs.append(net)
net = tf.concat(inputs, axis=-1)
tf.logging.info("inputs: {}".format(net))
if batch_norm:
net = tf.layers.batch_normalization(net, training=is_training)
net = add_hidden_layers(net, hidden_units, activation_fn, dropout,
is_training, batch_norm, 'DNN')
with tf.variable_scope('logits') as logits_scope:
logits = fc(net, head.logits_dimension, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return head.create_estimator_spec(
features=features,
mode=mode,
labels=labels,
optimizer=optimizers.get_optimizer_instance(
optimizer, learning_rate=_LEARNING_RATE),
logits=logits)
def get_cin_logits(features, feature_columns, shared_feature_vectors, units,
is_training, extra_options):
with tf.variable_scope('cin'):
_check_cin_args(extra_options)
use_shared_embedding = extra_options['cin_use_shared_embedding']
use_project = extra_options['cin_use_project']
project_size = extra_options['cin_project_size']
hidden_feature_maps = extra_options['cin_hidden_feature_maps']
split_half = extra_options['cin_split_half']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
check_feature_dims(feature_vectors)
x = tf.stack(feature_vectors, axis=1) # [B, N, D]
y = _cin_layer(x, hidden_feature_maps, split_half,
reduce_sum=False) # [B, F]
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units=units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_cross_logits(
features,
feature_columns,
shared_feature_vectors,
units,
is_training,
extra_options):
with tf.variable_scope('cross'):
_check_cross_args(extra_options)
use_shared_embedding = extra_options['cross_use_shared_embedding']
use_project = extra_options['cross_use_project']
project_size = extra_options['cross_project_size']
num_layers = extra_options['cross_num_layers']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
x = tf.concat(feature_vectors, axis=1) # [B, T]
y = _cross_net(x, num_layers)
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units=units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_fm_logits(features, feature_columns, shared_feature_vectors, units,
is_training, extra_options):
assert units == 1, "FM units must be 1"
with tf.variable_scope('fm'):
_check_fm_args(extra_options)
use_shared_embedding = extra_options['fm_use_shared_embedding']
use_project = extra_options['fm_use_project']
project_size = extra_options['fm_project_size']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
y = _fm(feature_vectors, reduce_sum=True) # [B, 1]
with tf.variable_scope('logits') as logits_scope:
logits = y
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_wkfm_logits(
features,
feature_columns,
shared_feature_vectors,
units,
is_training,
extra_options):
with tf.variable_scope('wkfm'):
_check_wkfm_args(extra_options)
use_shared_embedding = extra_options['wkfm_use_shared_embedding']
use_project = extra_options['wkfm_use_project']
project_size = extra_options['wkfm_project_size']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
y = _wkfm(feature_vectors, reduce_sum=True) # [B, 1]
with tf.variable_scope('logits') as logits_scope:
# fc just for adding a bias
logits = fc(y, units=units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_ipnn_logits(
features,
feature_columns,
shared_feature_vectors,
units,
is_training,
extra_options):
with tf.variable_scope('ipnn'):
_check_ipnn_args(extra_options)
use_shared_embedding = extra_options['ipnn_use_shared_embedding']
use_project = extra_options['ipnn_use_project']
project_size = extra_options['ipnn_project_size']
hidden_units = extra_options['ipnn_hidden_units']
activation_fn = extra_options['ipnn_activation_fn']
dropout = extra_options['ipnn_dropout']
batch_norm = extra_options['ipnn_batch_norm']
layer_norm = extra_options['ipnn_layer_norm']
use_resnet = extra_options['ipnn_use_resnet']
use_densenet = extra_options['ipnn_use_densenet']
unordered_inner_product = extra_options['ipnn_unordered_inner_product']
concat_project = extra_options['ipnn_concat_project']
leaky_relu_alpha = extra_options['leaky_relu_alpha']
swish_beta = extra_options['swish_beta']
activation_fn = get_activation_fn(activation_fn=activation_fn,
leaky_relu_alpha=leaky_relu_alpha,
swish_beta=swish_beta)
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
project_feature_vectors = None
if use_project:
project_feature_vectors = project(feature_vectors, project_size)
y = _ipnn(feature_vectors=feature_vectors,
project_feature_vectors=project_feature_vectors,
use_project=use_project,
units=units,
hidden_units=hidden_units,
activation_fn=activation_fn,
dropout=dropout,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet,
is_training=is_training,
unordered_inner_product=unordered_inner_product,
concat_project=concat_project)
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units=units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_linear_logits(features,
linear_feature_columns,
units,
linear_sparse_combiner,
scope):
logit_fn = linear.linear_logit_fn_builder(
units=units,
feature_columns=linear_feature_columns,
sparse_combiner=linear_sparse_combiner)
logits = logit_fn(features=features)
add_hidden_layer_summary(logits, scope.name)
return logits
def get_ccpm_logits(features, feature_columns, shared_feature_vectors, units,
is_training, extra_options):
with tf.variable_scope('ccpm'):
_check_ccpm_args(extra_options)
use_shared_embedding = extra_options['ccpm_use_shared_embedding']
use_project = extra_options['ccpm_use_project']
project_size = extra_options['ccpm_project_size']
hidden_units = extra_options['ccpm_hidden_units']
activation_fn = extra_options['ccpm_activation_fn']
dropout = extra_options['ccpm_dropout']
batch_norm = extra_options['ccpm_batch_norm']
layer_norm = extra_options['ccpm_layer_norm']
use_resnet = extra_options['ccpm_use_resnet']
use_densenet = extra_options['ccpm_use_densenet']
kernel_sizes = extra_options['ccpm_kernel_sizes']
filter_nums = extra_options['ccpm_filter_nums']
leaky_relu_alpha = extra_options['leaky_relu_alpha']
swish_beta = extra_options['swish_beta']
activation_fn = get_activation_fn(activation_fn=activation_fn,
leaky_relu_alpha=leaky_relu_alpha,
swish_beta=swish_beta)
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
y = _build_ccpm_model(feature_vectors=feature_vectors,
kernel_sizes=kernel_sizes,
filter_nums=filter_nums,
hidden_units=hidden_units,
activation_fn=activation_fn,
dropout=dropout,
is_training=is_training,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet)
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units=units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_nfm_logits(features, feature_columns, shared_feature_vectors, units,
is_training, extra_options):
with tf.variable_scope('nfm'):
_check_nfm_args(extra_options)
use_shared_embedding = extra_options['nfm_use_shared_embedding']
use_project = extra_options['nfm_use_project']
project_size = extra_options['nfm_project_size']
hidden_units = extra_options['nfm_hidden_units']
activation_fn = extra_options['nfm_activation_fn']
dropout = extra_options['nfm_dropout']
batch_norm = extra_options['nfm_batch_norm']
layer_norm = extra_options['nfm_layer_norm']
use_resnet = extra_options['nfm_use_resnet']
use_densenet = extra_options['nfm_use_densenet']
leaky_relu_alpha = extra_options['leaky_relu_alpha']
swish_beta = extra_options['swish_beta']
activation_fn = get_activation_fn(activation_fn=activation_fn,
leaky_relu_alpha=leaky_relu_alpha,
swish_beta=swish_beta)
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
# Neural FM
y = _fm(feature_vectors, reduce_sum=False)
y = add_hidden_layers(y,
hidden_units=hidden_units,
activation_fn=activation_fn,
dropout=dropout,
is_training=is_training,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet,
scope='hidden_layers')
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_autoint_logits(
features,
feature_columns,
shared_feature_vectors,
units,
is_training,
extra_options):
with tf.variable_scope('autoint'):
_check_autoint_args(extra_options)
use_shared_embedding = extra_options['autoint_use_shared_embedding']
use_project = extra_options['autoint_use_project']
project_size = extra_options['autoint_project_size']
size_per_head = extra_options['autoint_size_per_head']
num_heads = extra_options['autoint_num_heads']
num_blocks = extra_options['autoint_num_blocks']
dropout = extra_options['autoint_dropout']
has_residual = extra_options['autoint_has_residual']
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
check_feature_dims(feature_vectors)
x = tf.stack(feature_vectors, axis=1) # [B, N, D]
y = _autoint(x,
num_blocks=num_blocks,
num_units=size_per_head*num_heads,
num_heads=num_heads,
dropout=dropout,
is_training=is_training,
has_residual=has_residual)
tf.logging.info("autoint output = {}".format(y))
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def get_fibinet_logits(
features,
feature_columns,
shared_feature_vectors,
units,
is_training,
extra_options):
with tf.variable_scope('fibinet'):
_check_fibinet_args(extra_options)
use_shared_embedding = extra_options['fibinet_use_shared_embedding']
use_project = extra_options['fibinet_use_project']
project_size = extra_options['fibinet_project_size']
hidden_units = extra_options['fibinet_hidden_units']
activation_fn = extra_options['fibinet_activation_fn']
dropout = extra_options['fibinet_dropout']
batch_norm = extra_options['fibinet_batch_norm']
layer_norm = extra_options['fibinet_layer_norm']
use_resnet = extra_options['fibinet_use_resnet']
use_densenet = extra_options['fibinet_use_densenet']
use_se = extra_options['fibinet_use_se']
use_deep = extra_options['fibinet_use_deep']
interaction_type = extra_options['fibinet_interaction_type']
se_interaction_type = extra_options['fibinet_se_interaction_type']
se_use_shared_embedding = extra_options['fibinet_se_use_shared_embedding']
leaky_relu_alpha = extra_options['leaky_relu_alpha']
swish_beta = extra_options['swish_beta']
activation_fn = get_activation_fn(activation_fn=activation_fn,
leaky_relu_alpha=leaky_relu_alpha,
swish_beta=swish_beta)
if not use_shared_embedding:
feature_vectors = get_feature_vectors(features, feature_columns)
else:
feature_vectors = shared_feature_vectors
if use_project:
feature_vectors = project(feature_vectors, project_size)
y = shallow_fibinet(features=features,
feature_columns=feature_columns,
shared_feature_vectors=feature_vectors,
se_use_shared_embedding=se_use_shared_embedding,
use_project=use_project,
project_size=project_size,
interaction_type=interaction_type,
se_interaction_type=se_interaction_type,
use_se=use_se) # [B, -1]
if use_deep:
y = add_hidden_layers(y,
hidden_units=hidden_units,
activation_fn=activation_fn,
dropout=dropout,
is_training=is_training,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet,
scope='hidden_layers')
with tf.variable_scope('logits') as logits_scope:
logits = fc(y, units, name=logits_scope)
add_hidden_layer_summary(logits, logits_scope.name)
else:
assert units == 1, "shallow_fibinet's units must be 1"
with tf.variable_scope('logits') as logits_scope:
logits = tf.reduce_sum(y, axis=-1, keepdims=True) # [B, 1]
add_hidden_layer_summary(logits, logits_scope.name)
return logits
def _model_fn(features, labels, mode, config):
head = head_lib._binary_logistic_or_multi_class_head( # pylint: disable=protected-access
n_classes, weight_column, label_vocabulary, loss_reduction,
loss_fn)
is_training = (mode == tf.estimator.ModeKeys.TRAIN)
net_dssm1 = tf.feature_column.input_layer(features, dssm1_columns)
net_dssm2 = tf.feature_column.input_layer(features, dssm2_columns)
tf.logging.info("net_dssm1: {}".format(net_dssm1))
tf.logging.info("net_dssm2: {}".format(net_dssm2))
real_activation_fn = get_activation_fn(
activation_fn=activation_fn,
leaky_relu_alpha=leaky_relu_alpha,
swish_beta=swish_beta)
net_dssm1 = add_hidden_layers(inputs=net_dssm1,
hidden_units=hidden_units,
activation_fn=real_activation_fn,
dropout=dropout,
is_training=is_training,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet,
scope='dssm1')
net_dssm2 = add_hidden_layers(inputs=net_dssm2,
hidden_units=hidden_units,
activation_fn=real_activation_fn,
dropout=dropout,
is_training=is_training,
batch_norm=batch_norm,
layer_norm=layer_norm,
use_resnet=use_resnet,
use_densenet=use_densenet,
scope='dssm2')
with tf.variable_scope('logits') as logits_scope:
if dssm_mode == 'dot':
logits = tf.reduce_sum(net_dssm1 * net_dssm2,
-1,
keepdims=True)
elif dssm_mode == 'concat':
logits = tf.concat([net_dssm1, net_dssm2], axis=1)
logits = tf.layers.dense(logits, units=1, activation=None)
elif dssm_mode == 'cosine':
logits = tf.reduce_sum(net_dssm1 * net_dssm2,
-1,
keepdims=True)
norm1 = tf.norm(net_dssm1, axis=1, keepdims=True)
norm2 = tf.norm(net_dssm2, axis=1, keepdims=True)
logits = logits / (norm1 * norm2)
else:
raise ValueError(
"unknown dssm mode '{}'".format(dssm_mode))
add_hidden_layer_summary(logits, logits_scope.name)
tf.logging.info("logits = {}".format(logits))
return head.create_estimator_spec(
features=features,
mode=mode,
labels=labels,
optimizer=optimizers.get_optimizer_instance(
optimizer, learning_rate=_LEARNING_RATE),
logits=logits)