def _get_default_head(params, weights_name, output_type, name=None):
"""Creates a default head based on a type of a problem."""
if output_type == ModelBuilderOutputType.MODEL_FN_OPS:
if params.regression:
return head_lib.regression_head(
weight_column_name=weights_name,
label_dimension=params.num_outputs,
enable_centered_bias=False,
head_name=name)
else:
return head_lib.multi_class_head(
params.num_classes,
weight_column_name=weights_name,
enable_centered_bias=False,
head_name=name)
else:
if params.regression:
return core_head_lib._regression_head( # pylint:disable=protected-access
weight_column=weights_name,
label_dimension=params.num_outputs,
name=name,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
else:
return core_head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=params.num_classes,
weight_column=weights_name,
name=name,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
def testWithFeatureColumns(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3, loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(
num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(
hparams.fill(),
head=head_fn,
feature_columns=[core_feature_column.numeric_column('x')])
iris = base.load_iris()
data = {'x': iris.data.astype(np.float32)}
labels = iris.target.astype(np.int32)
input_fn = numpy_io.numpy_input_fn(
x=data, y=labels, batch_size=150, num_epochs=None, shuffle=False)
est.train(input_fn=input_fn, steps=100)
res = est.evaluate(input_fn=input_fn, steps=1)
self.assertEqual(1.0, res['accuracy'])
self.assertAllClose(0.55144483, res['loss'])
def testTrainEvaluateInferDoesNotThrowErrorForClassifier(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(hparams.fill(),
head=head_fn)
input_fn, predict_input_fn = _get_classification_input_fns()
est.train(input_fn=input_fn, steps=100)
res = est.evaluate(input_fn=input_fn, steps=1)
self.assertEqual(1.0, res['accuracy'])
self.assertAllClose(0.55144483, res['loss'])
predictions = list(est.predict(input_fn=predict_input_fn))
self.assertAllClose([[0.576117, 0.211942, 0.211942]],
[pred['probabilities'] for pred in predictions])
def testTrainEvaluateInferDoesNotThrowErrorForClassifier(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3, loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(
num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(hparams.fill(), head=head_fn)
input_fn, predict_input_fn = _get_classification_input_fns()
est.train(input_fn=input_fn, steps=100)
res = est.evaluate(input_fn=input_fn, steps=1)
self.assertEqual(1.0, res['accuracy'])
self.assertAllClose(0.55144483, res['loss'])
predictions = list(est.predict(input_fn=predict_input_fn))
self.assertAllClose([[0.576117, 0.211942, 0.211942]],
[pred['probabilities'] for pred in predictions])
def _get_default_head(params, weights_name, output_type, name=None):
"""Creates a default head based on a type of a problem."""
if output_type == ModelBuilderOutputType.MODEL_FN_OPS:
if params.regression:
return head_lib.regression_head(weight_column_name=weights_name,
label_dimension=params.num_outputs,
enable_centered_bias=False,
head_name=name)
else:
return head_lib.multi_class_head(params.num_classes,
weight_column_name=weights_name,
enable_centered_bias=False,
head_name=name)
else:
if params.regression:
return core_head_lib._regression_head( # pylint:disable=protected-access
weight_column=weights_name,
label_dimension=params.num_outputs,
name=name,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
else:
return core_head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=params.num_classes,
weight_column=weights_name,
name=name,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
def testWithFeatureColumns(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(
hparams.fill(),
head=head_fn,
feature_columns=[core_feature_column.numeric_column('x')])
iris = base.load_iris()
data = {'x': iris.data.astype(np.float32)}
labels = iris.target.astype(np.int32)
input_fn = numpy_io.numpy_input_fn(x=data,
y=labels,
batch_size=150,
num_epochs=None,
shuffle=False)
est.train(input_fn=input_fn, steps=100)
res = est.evaluate(input_fn=input_fn, steps=1)
self.assertEqual(1.0, res['accuracy'])
self.assertAllClose(0.55144483, res['loss'])
def __init__(self,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Ftrl',
config=None,
loss_reduction=losses.Reduction.SUM):
"""Initializes a BaselineClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
It must be greater than 1. Note: Class labels are integers representing
the class index (i.e. values from 0 to n_classes-1). For arbitrary
label values (e.g. string labels), convert to class indices first.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It will be multiplied by the loss of the example.
label_vocabulary: Optional list of strings with size `[n_classes]`
defining the label vocabulary. Only supported for `n_classes` > 2.
optimizer: String, `tf.Optimizer` object, or callable that creates the
optimizer to use for training. If not specified, will use
`FtrlOptimizer` with a default learning rate of 0.3.
config: `RunConfig` object to configure the runtime settings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM`.
Returns:
A `BaselineClassifier` estimator.
Raises:
ValueError: If `n_classes` < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
return _baseline_model_fn(features=features,
labels=labels,
mode=mode,
head=head,
optimizer=optimizer,
weight_column=weight_column,
config=config)
super(BaselineClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
hidden_units,
feature_columns,
model_dir=None,
n_classes=2,
weight_feature_key=None,
optimizer='Adagrad',
activation_fn=nn.relu,
dropout=None,
input_layer_partitioner=None,
config=None):
"""Initializes a `DNNClassifier` instance.
Args:
hidden_units: Iterable of number hidden units per layer. All layers are
fully connected. Ex. `[64, 32]` means first layer has 64 nodes and
second one has 32.
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `_FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_feature_key: A string defining feature column name representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
optimizer: An instance of `tf.Optimizer` used to train the model. If
`None`, will use an Adagrad optimizer.
activation_fn: Activation function applied to each layer. If `None`, will
use `tf.nn.relu`.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_feature_key=weight_feature_key)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes, weight_feature_key=weight_feature_key)
def _model_fn(features, labels, mode, config):
return _dnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
hidden_units=hidden_units,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
activation_fn=activation_fn,
dropout=dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(self,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Ftrl',
config=None,
loss_reduction=losses.Reduction.SUM):
"""Initializes a BaselineClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
It must be greater than 1. Note: Class labels are integers representing
the class index (i.e. values from 0 to n_classes-1). For arbitrary
label values (e.g. string labels), convert to class indices first.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It will be multiplied by the loss of the example.
label_vocabulary: Optional list of strings with size `[n_classes]`
defining the label vocabulary. Only supported for `n_classes` > 2.
optimizer: String, `tf.Optimizer` object, or callable that creates the
optimizer to use for training. If not specified, will use
`FtrlOptimizer` with a default learning rate of 0.3.
config: `RunConfig` object to configure the runtime settings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM`.
Returns:
A `BaselineClassifier` estimator.
Raises:
ValueError: If `n_classes` < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes, weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
return _baseline_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
optimizer=optimizer,
weight_column=weight_column,
config=config)
super(BaselineClassifier, self).__init__(
model_fn=_model_fn,
model_dir=model_dir,
config=config)
def multi_class_head(n_classes,
weight_column=None,
label_vocabulary=None,
loss_reduction=losses.Reduction.SUM,
name=None):
"""Creates a `_Head` for multi class classification.
Uses `sparse_softmax_cross_entropy` loss.
The head expects `logits` with shape `[D0, D1, ... DN, n_classes]`.
In many applications, the shape is `[batch_size, n_classes]`.
`labels` must be a dense `Tensor` with shape matching `logits`, namely
`[D0, D1, ... DN, 1]`. If `label_vocabulary` given, `labels` must be a string
`Tensor` with values from the vocabulary. If `label_vocabulary` is not given,
`labels` must be an integer `Tensor` with values specifying the class index.
If `weight_column` is specified, weights must be of shape
`[D0, D1, ... DN]`, or `[D0, D1, ... DN, 1]`.
The loss is the weighted sum over the input dimensions. Namely, if the input
labels have shape `[batch_size, 1]`, the loss is the weighted sum over
`batch_size`.
Args:
n_classes: Number of classes, must be greater than 2 (for 2 classes, use
`binary_classification_head`).
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
label_vocabulary: A list or tuple of strings representing possible label
values. If it is not given, that means labels are already encoded as an
integer within [0, n_classes). If given, labels must be of string type and
have any value in `label_vocabulary`. Note that errors will be raised if
`label_vocabulary` is not provided but labels are strings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how to
reduce training loss over batch. Defaults to `SUM`.
name: name of the head. If provided, summary and metrics keys will be
suffixed by `"/" + name`. Also used as `name_scope` when creating ops.
Returns:
An instance of `_Head` for multi class classification.
Raises:
ValueError: if `n_classes`, `label_vocabulary` or `loss_reduction` is
invalid.
"""
return head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction,
name=name)
def multi_class_head(n_classes,
weight_column=None,
label_vocabulary=None,
loss_reduction=losses.Reduction.SUM,
name=None):
"""Creates a `_Head` for multi class classification.
Uses `sparse_softmax_cross_entropy` loss.
The head expects `logits` with shape `[D0, D1, ... DN, n_classes]`.
In many applications, the shape is `[batch_size, n_classes]`.
`labels` must be a dense `Tensor` with shape matching `logits`, namely
`[D0, D1, ... DN, 1]`. If `label_vocabulary` given, `labels` must be a string
`Tensor` with values from the vocabulary. If `label_vocabulary` is not given,
`labels` must be an integer `Tensor` with values specifying the class index.
If `weight_column` is specified, weights must be of shape
`[D0, D1, ... DN]`, or `[D0, D1, ... DN, 1]`.
The loss is the weighted sum over the input dimensions. Namely, if the input
labels have shape `[batch_size, 1]`, the loss is the weighted sum over
`batch_size`.
Args:
n_classes: Number of classes, must be greater than 2 (for 2 classes, use
`binary_classification_head`).
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
label_vocabulary: A list or tuple of strings representing possible label
values. If it is not given, that means labels are already encoded as an
integer within [0, n_classes). If given, labels must be of string type and
have any value in `label_vocabulary`. Note that errors will be raised if
`label_vocabulary` is not provided but labels are strings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how to
reduce training loss over batch. Defaults to `SUM`.
name: name of the head. If provided, summary and metrics keys will be
suffixed by `"/" + name`. Also used as `name_scope` when creating ops.
Returns:
An instance of `_Head` for multi class classification.
Raises:
ValueError: if `n_classes`, `label_vocabulary` or `loss_reduction` is
invalid.
"""
return head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction,
name=name)
def __init__(self,
feature_columns,
model_dir=None,
n_classes=2,
weight_feature_key=None,
optimizer=None,
config=None,
partitioner=None):
"""Construct a `LinearClassifier` estimator object.
Args:
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
Note that class labels are integers representing the class index (i.e.
values from 0 to n_classes-1). For arbitrary label values (e.g. string
labels), convert to class indices first.
weight_feature_key: A string defining feature column name representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
optimizer: The optimizer used to train the model. If specified, it should
be either an instance of `tf.Optimizer` or the SDCAOptimizer. If `None`,
the Ftrl optimizer will be used.
config: `RunConfig` object to configure the runtime settings.
partitioner: Optional. Partitioner for input layer.
Returns:
A `LinearClassifier` estimator.
Raises:
ValueError: if n_classes < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_feature_key=weight_feature_key)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_feature_key=weight_feature_key)
super(LinearClassifier, self).__init__(model_fn=_linear_model_fn,
model_dir=model_dir,
config=config,
params={
'head': head,
'feature_columns':
feature_columns,
'optimizer': optimizer,
'partitioner': partitioner,
})
def __init__(self,
num_workers,
model_dir=None,
linear_feature_columns=None,
dnn_feature_columns=None,
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None):
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = (
list(linear_feature_columns) + list(dnn_feature_columns))
if not self._feature_columns:
raise ValueError('Either linear_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary)
self.optimizer = self.create_optimizer(num_workers, linear_feature_columns)
def _model_fn(features, labels, mode, config):
return _dnn_linear_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
dnn_feature_columns=dnn_feature_columns,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config,
num_workers=num_workers,
opt=self.optimizer)
super(DNNLinearCombinedClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(
self,
hidden_units,
linear_feature_columns,
dnn_feature_columns,
fm_feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
activation_fn=nn.relu,
dropout=None,
input_layer_partitioner=None,
config=None,
warm_start_from=None,
loss_reduction=losses.Reduction.SUM,
):
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
"""Call the defined shared _dnn_model_fn."""
return _dfm_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
hidden_units=hidden_units,
dnn_feature_columns=dnn_feature_columns,
linear_feature_columns=linear_feature_columns,
fm_feature_columns=fm_feature_columns,
optimizer=optimizer,
activation_fn=activation_fn,
dropout=dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DeepFMClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config,
warm_start_from=warm_start_from)
def core_multiclass_head(n_classes):
"""Core head for multiclass problems."""
def loss_fn(labels, logits):
result = losses.per_example_maxent_loss(
labels=labels, logits=logits, weights=None, num_classes=n_classes)
return result[0]
# pylint:disable=protected-access
head_fn = core_head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=n_classes,
loss_fn=loss_fn,
loss_reduction=core_losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
# pylint:enable=protected-access
return head_fn
def core_multiclass_head(n_classes):
"""Core head for multiclass problems."""
def loss_fn(labels, logits):
result = losses.per_example_maxent_loss(labels=labels,
logits=logits,
weights=None,
num_classes=n_classes)
return result[0]
# pylint:disable=protected-access
head_fn = core_head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=n_classes,
loss_fn=loss_fn,
loss_reduction=core_losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
# pylint:enable=protected-access
return head_fn
def testEarlyStopping(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3, loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(
num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(
hparams.fill(),
head=head_fn,
# Set a crazy threshold - 30% loss change.
early_stopping_loss_threshold=0.3,
early_stopping_rounds=2)
input_fn, _ = _get_classification_input_fns()
est.train(input_fn=input_fn, steps=100)
# We stopped early.
self._assert_checkpoint(est.model_dir, global_step=8)
def testEarlyStopping(self):
head_fn = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes=3,
loss_reduction=losses.Reduction.SUM_OVER_NONZERO_WEIGHTS)
hparams = tensor_forest.ForestHParams(num_trees=3,
max_nodes=1000,
num_classes=3,
num_features=4,
split_after_samples=20,
inference_tree_paths=True)
est = random_forest.CoreTensorForestEstimator(
hparams.fill(),
head=head_fn,
# Set a crazy threshold - 30% loss change.
early_stopping_loss_threshold=0.3,
early_stopping_rounds=2)
input_fn, _ = _get_classification_input_fns()
est.train(input_fn=input_fn, steps=100)
# We stopped early.
self._assert_checkpoint(est.model_dir, global_step=8)
def multi_class_head(n_classes,
weight_column=None,
label_vocabulary=None,
name=None):
"""Creates a `_Head` for multi class classification.
Uses `sparse_softmax_cross_entropy` loss.
This head expects to be fed integer labels specifying the class index.
Args:
n_classes: Number of classes, must be greater than 2 (for 2 classes, use
`binary_classification_head`).
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
label_vocabulary: A list of strings represents possible label values. If it
is not given, that means labels are already encoded as integer within
[0, n_classes). If given, labels must be string type and have any value in
`label_vocabulary`. Also there will be errors if vocabulary is not
provided and labels are string.
name: name of the head. If provided, summary and metrics keys will be
suffixed by `"/" + name`.
Returns:
An instance of `_Head` for multi class classification.
Raises:
ValueError: if `n_classes`, `metric_class_ids` or `label_keys` is invalid.
"""
return head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
name=name)
def __init__(self,
model_collections,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
input_layer_partitioner=None,
config=None):
"""Initializes a `DNNClassifier` instance.
Args:
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `_FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Defaults
to Adagrad optimizer.
activation_fn: Activation function applied to each layer. If `None`, will
use `tf.nn.relu`.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
"""
if not model_collections:
raise ValueError(
'Empty model collections, must fill DNN model instance.')
assert isinstance(
model_collections,
(list, tuple)), "model_collections must be a list or tuple"
for model in model_collections:
if not isinstance(model, DNN):
raise ValueError(
"model_collections element must be an instance of class DNN"
)
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _dnn_model_fn(features,
labels,
mode,
head,
optimizer='Adagrad',
input_layer_partitioner=None,
config=None):
"""Deep Neural Net model_fn.
Args:
features: dict of `Tensor`.
labels: `Tensor` of shape [batch_size, 1] or [batch_size] labels of
dtype `int32` or `int64` in the range `[0, n_classes)`.
mode: Defines whether this is training, evaluation or prediction.
See `ModeKeys`.
head: A `head_lib._Head` instance.
hidden_units: Iterable of integer number of hidden units per layer.
feature_columns: Iterable of `feature_column._FeatureColumn` model inputs.
optimizer: String, `tf.Optimizer` object, or callable that creates the
optimizer to use for training. If not specified, will use the Adagrad
optimizer with a default learning rate of 0.05.
activation_fn: Activation function applied to each layer.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
Returns:
predictions: A dict of `Tensor` objects.
loss: A scalar containing the loss of the step.
train_op: The op for training.
Raises:
ValueError: If features has the wrong type.
"""
if not isinstance(features, dict):
raise ValueError(
'features should be a dictionary of `Tensor`s. '
'Given type: {}'.format(type(features)))
optimizer = get_optimizer_instance(optimizer, learning_rate=0.05)
num_ps_replicas = config.num_ps_replicas if config else 0
partitioner = tf.compat.v1.min_max_variable_partitioner(
max_partitions=num_ps_replicas)
with tf.compat.v1.variable_scope('dnn',
values=tuple(
six.itervalues(features)),
partitioner=partitioner):
input_layer_partitioner = input_layer_partitioner or (
tf.compat.v1.min_max_variable_partitioner(
max_partitions=num_ps_replicas,
min_slice_size=64 << 20))
# unit is num_classes, shape(batch_size, num_classes)
logits = []
for idx, m in enumerate(model_collections):
logits.append(
_dnn_logit_fn(features, mode, idx + 1,
head.logits_dimension, m.hidden_units,
m.connected_layers, feature_columns,
input_layer_partitioner))
logits = tf.add_n(
logits
) # add logit layer is same with concactenate the layer before logit layer
def _train_op_fn(loss):
"""Returns the op to optimize the loss."""
return optimizer.minimize(
loss, global_step=tf.compat.v1.train.get_global_step())
return head.create_estimator_spec(features=features,
mode=mode,
labels=labels,
train_op_fn=_train_op_fn,
logits=logits)
def _model_fn(features, labels, mode, config):
return _dnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
optimizer=optimizer,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(MultiDNNClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
model_type=None,
model_dir=None,
with_cnn=False,
cnn_optimizer='Adagrad',
linear_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_connected_mode=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None):
"""Initializes a WideDeepCombinedClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
linear_feature_columns: An iterable containing all the feature columns
used by linear part of the model. All items in the set must be
instances of classes derived from `FeatureColumn`.
linear_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the linear part of the model. Defaults to FTRL optimizer.
dnn_feature_columns: An iterable containing all the feature columns used
by deep part of the model. All items in the set must be instances of
classes derived from `FeatureColumn`.
dnn_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the deep part of the model. Defaults to Adagrad optimizer.
dnn_hidden_units: List of hidden units per layer. All layers are fully
connected.
dnn_activation_fn: Activation function applied to each layer. If None,
will use `tf.nn.relu`.
dnn_dropout: When not None, the probability we will drop out
a given coordinate.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
input_layer_partitioner: Partitioner for input layer. Defaults to
`min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: RunConfig object to configure the runtime settings.
Raises:
ValueError: If both linear_feature_columns and dnn_features_columns are
empty at the same time.
"""
if not linear_feature_columns and not dnn_feature_columns:
raise ValueError('Either linear_feature_columns or dnn_feature_columns must be defined.')
if model_type is None:
raise ValueError("Model type must be defined. one of `wide_deep`, `deep`, `wide_deep`.")
else:
assert model_type in {'wide_deep', 'deep', 'wide_deep'}, (
"Invalid model type, must be one of `wide_deep`, `deep`, `wide_deep`.")
if model_type == 'wide_deep':
if not linear_feature_columns:
raise ValueError('Linear_feature_columns must be defined for wide_deep model.')
elif model_type == 'deep':
if not dnn_feature_columns:
raise ValueError('Dnn_feature_columns must be defined for deep model.')
if dnn_feature_columns and not dnn_hidden_units:
raise ValueError('dnn_hidden_units must be defined when dnn_feature_columns is specified.')
if n_classes == 2:
# units = 1
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
# units = n_classes
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _model_fn(features, labels, mode, config):
return _wide_deep_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
model_type=model_type,
with_cnn=with_cnn,
cnn_optimizer=cnn_optimizer,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_connected_mode=dnn_connected_mode,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(WideAndDeepClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(self,
sequence_feature_columns,
context_feature_columns=None,
num_units=None,
cell_type=USE_DEFAULT,
rnn_cell_fn=None,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
loss_reduction=losses.Reduction.SUM_OVER_BATCH_SIZE,
input_layer_partitioner=None,
config=None):
"""Initializes a `RNNClassifier` instance.
Args:
sequence_feature_columns: An iterable containing the `FeatureColumn`s
that represent sequential input. All items in the set should either be
sequence columns (e.g. `sequence_numeric_column`) or constructed from
one (e.g. `embedding_column` with `sequence_categorical_column_*` as
input).
context_feature_columns: An iterable containing the `FeatureColumn`s
for contextual input. The data represented by these columns will be
replicated and given to the RNN at each timestep. These columns must be
instances of classes derived from `_DenseColumn` such as
`numeric_column`, not the sequential variants.
num_units: Iterable of integer number of hidden units per RNN layer. If
set, `cell_type` must also be specified and `rnn_cell_fn` must be
`None`.
cell_type: A subclass of `tf.nn.rnn_cell.RNNCell` or a string specifying
the cell type. Supported strings are: `'basic_rnn'`, `'lstm'`, and
`'gru'`. If set, `num_units` must also be specified and `rnn_cell_fn`
must be `None`.
rnn_cell_fn: A function with one argument, a `tf.estimator.ModeKeys`, and
returns an object of type `tf.nn.rnn_cell.RNNCell` that will be used to
construct the RNN. If set, `num_units` and `cell_type` cannot be set.
This is for advanced users who need additional customization beyond
`num_units` and `cell_type`. Note that `tf.nn.rnn_cell.MultiRNNCell` is
needed for stacked RNNs.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` or string specifying optimizer
type. Defaults to Adagrad optimizer.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM_OVER_BATCH_SIZE`.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
Raises:
ValueError: If `num_units`, `cell_type`, and `rnn_cell_fn` are not
compatible.
"""
rnn_cell_fn = _assert_rnn_cell_fn(rnn_cell_fn, num_units, cell_type)
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
return _rnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
rnn_cell_fn=rnn_cell_fn,
sequence_feature_columns=tuple(sequence_feature_columns or []),
context_feature_columns=tuple(context_feature_columns or []),
optimizer=optimizer,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(RNNClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(self,
model_dir=None,
fm_first_feature_columns=None,
fm_second_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
embedding_size=None,
field_size=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None,
warm_start_from=None,
loss_reduction=losses.Reduction.SUM):
fm_first_feature_columns = fm_first_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
fm_second_feature_columns = fm_second_feature_columns or []
self._feature_columns = (list(fm_first_feature_columns) +
list(dnn_feature_columns) +
list(fm_second_feature_columns))
if not self._feature_columns:
raise ValueError(
'Either fm_first_feature_columns or fm_second_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss(
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
"""Call the _deepfm_model_fn."""
return _deepfm_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
fm_first_feature_columns=fm_first_feature_columns,
fm_second_feature_columns=fm_second_feature_columns,
embedding_size=embedding_size,
field_size=field_size,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DeepFM, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config,
warm_start_from=warm_start_from)
def __init__(self,
hidden_units,
feature_columns,
model_dir=None,
n_classes=2,
weight_feature_key=None,
optimizer='Adagrad',
activation_fn=nn.relu,
dropout=None,
input_layer_partitioner=None,
config=None):
"""Initializes a `DNNClassifier` instance.
Args:
hidden_units: Iterable of number hidden units per layer. All layers are
fully connected. Ex. `[64, 32]` means first layer has 64 nodes and
second one has 32.
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `_FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_feature_key: A string defining feature column name representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
optimizer: An instance of `tf.Optimizer` used to train the model. If
`None`, will use an Adagrad optimizer.
activation_fn: Activation function applied to each layer. If `None`, will
use `tf.nn.relu`.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_feature_key=weight_feature_key)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_feature_key=weight_feature_key)
def _model_fn(features, labels, mode, config):
return _dnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
hidden_units=hidden_units,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
activation_fn=activation_fn,
dropout=dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
hidden_units,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
activation_fn=nn.relu,
dropout=None,
input_layer_partitioner=None,
config=None):
"""Initializes a `DNNClassifier` instance.
Args:
hidden_units: Iterable of number hidden units per layer. All layers are
fully connected. Ex. `[64, 32]` means first layer has 64 nodes and
second one has 32.
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `_FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Defaults
to Adagrad optimizer.
activation_fn: Activation function applied to each layer. If `None`, will
use `tf.nn.relu`.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes, weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _model_fn(features, labels, mode, config):
return _dnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
hidden_units=hidden_units,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
activation_fn=activation_fn,
dropout=dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(self,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Ftrl',
config=None,
partitioner=None,
warm_start_from=None):
"""Construct a `LinearClassifier` estimator object.
Args:
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
Note that class labels are integers representing the class index (i.e.
values from 0 to n_classes-1). For arbitrary label values (e.g. string
labels), convert to class indices first.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Defaults
to FTRL optimizer.
config: `RunConfig` object to configure the runtime settings.
partitioner: Optional. Partitioner for input layer.
warm_start_from: A string filepath to a checkpoint to warm-start from, or
a `WarmStartSettings` object to fully configure warm-starting. If the
string filepath is provided instead of a `WarmStartSettings`, then all
weights and biases are warm-started, and it is assumed that vocabularies
and Tensor names are unchanged.
Returns:
A `LinearClassifier` estimator.
Raises:
ValueError: if n_classes < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes, weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _model_fn(features, labels, mode, config):
"""Call the defined shared _linear_model_fn and possibly warm-start."""
estimator_spec = _linear_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
partitioner=partitioner,
config=config)
# pylint: disable=protected-access
warm_start_settings = warm_starting_util._get_default_warm_start_settings(
warm_start_from)
if warm_start_settings:
warm_starting_util._warm_start(warm_start_settings)
# pylint: enable=protected-access
return estimator_spec
super(LinearClassifier, self).__init__(
model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
model_dir=None,
linear_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
input_layer_partitioner=None,
config=None):
"""Initializes a DNNLinearCombinedClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
linear_feature_columns: An iterable containing all the feature columns
used by linear part of the model. All items in the set must be
instances of classes derived from `FeatureColumn`.
linear_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the linear part of the model. Defaults to FTRL optimizer.
dnn_feature_columns: An iterable containing all the feature columns used
by deep part of the model. All items in the set must be instances of
classes derived from `FeatureColumn`.
dnn_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the deep part of the model. Defaults to Adagrad optimizer.
dnn_hidden_units: List of hidden units per layer. All layers are fully
connected.
dnn_activation_fn: Activation function applied to each layer. If None,
will use `tf.nn.relu`.
dnn_dropout: When not None, the probability we will drop out
a given coordinate.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
input_layer_partitioner: Partitioner for input layer. Defaults to
`min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: RunConfig object to configure the runtime settings.
Raises:
ValueError: If both linear_feature_columns and dnn_features_columns are
empty at the same time.
"""
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = linear_feature_columns + dnn_feature_columns
if not self._feature_columns:
raise ValueError(
'Either linear_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss(
) # pylint: disable=protected-access
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes)
def _model_fn(features, labels, mode, config):
return _dnn_linear_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNLinearCombinedClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
model_dir=None,
linear_feature_columns=None,
linear_optimizer='Ftrl',
linear_sparse_combiner='sum',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
fm_embedding_size=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None,
warm_start_from=None,
loss_reduction=tf.losses.Reduction.SUM,
batch_norm=False):
"""
仿照wide_deep官方模型写法的deepFM模型，使用feature_column作特征处理。
* 不用将数据集转换为libsvm格式.
* 支持multi-hot类型特征，例如用户的历史消费物品序列.
* 暂不支持数值型特征的交叉.
【特征说明】
参考train.py.
1. 原始特征分为两种：
* 数值特征(dense)：没啥好说的
* 离散特征(sparse)：id类特征（vocabulary size特别大，使用特征hash），非id类特征（vocabulary size在几个到几百个之间，直接one-hot编码）
特别的，当多个不定长离散特征组成一个序列，直接编码就是multi-hot特征，非常常见。
2. 简要叙述官方wide_deep
原生的wide部分由于其实现方式，直接支持sparse tensor的输入；
而dnn部分只支持dense tensor输入。
这里直接copy官方写法，也分为两部分输入。
【DeepFM简要说明】
* 按理说，deepFM的一阶项和二阶项应该都是相同的输入，分为两部分是仿照wide_deep写法，能够直接复用linear model代码。
* 数值特征目前还不支持。
* 将离散特征直接传给linear_feature_columns，作为一阶项的输入。
* 将离散特征通过embedding变为dense tensor后传给dnn_feature_columns，作为二阶项和dnn的输入。
* 实质上，一阶项和二阶项的输入是相同的。
* 为了稍微利用数值特征，将数值特征分桶后用在了linear部分，而二阶项和dnn部分目前没有使用数值特征（todo）
"""
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = (list(linear_feature_columns) +
list(dnn_feature_columns))
if not self._feature_columns:
raise ValueError('empty columns.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
return _DeepFM_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
linear_sparse_combiner=linear_sparse_combiner,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
fm_embedding_size=fm_embedding_size,
input_layer_partitioner=input_layer_partitioner,
config=config,
batch_norm=batch_norm)
super(DeepFM, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config,
warm_start_from=warm_start_from)
def multi_class_head(n_classes,
weight_column=None,
label_vocabulary=None,
loss_reduction=losses.Reduction.SUM_OVER_BATCH_SIZE,
loss_fn=None,
name=None):
"""Creates a `_Head` for multi class classification.
Uses `sparse_softmax_cross_entropy` loss.
The head expects `logits` with shape `[D0, D1, ... DN, n_classes]`.
In many applications, the shape is `[batch_size, n_classes]`.
`labels` must be a dense `Tensor` with shape matching `logits`, namely
`[D0, D1, ... DN, 1]`. If `label_vocabulary` given, `labels` must be a string
`Tensor` with values from the vocabulary. If `label_vocabulary` is not given,
`labels` must be an integer `Tensor` with values specifying the class index.
If `weight_column` is specified, weights must be of shape
`[D0, D1, ... DN]`, or `[D0, D1, ... DN, 1]`.
The loss is the weighted sum over the input dimensions. Namely, if the input
labels have shape `[batch_size, 1]`, the loss is the weighted sum over
`batch_size`.
Also supports custom `loss_fn`. `loss_fn` takes `(labels, logits)` or
`(labels, logits, features)` as arguments and returns unreduced loss with
shape `[D0, D1, ... DN, 1]`. `loss_fn` must support integer `labels` with
shape `[D0, D1, ... DN, 1]`. Namely, the head applies `label_vocabulary` to
the input labels before passing them to `loss_fn`.
The head can be used with a canned estimator. Example:
```python
my_head = tf.contrib.estimator.multi_class_head(n_classes=3)
my_estimator = tf.contrib.estimator.DNNEstimator(
head=my_head,
hidden_units=...,
feature_columns=...)
```
It can also be used with a custom `model_fn`. Example:
```python
def _my_model_fn(features, labels, mode):
my_head = tf.contrib.estimator.multi_class_head(n_classes=3)
logits = tf.keras.Model(...)(features)
return my_head.create_estimator_spec(
features=features,
mode=mode,
labels=labels,
optimizer=tf.AdagradOptimizer(learning_rate=0.1),
logits=logits)
my_estimator = tf.estimator.Estimator(model_fn=_my_model_fn)
```
Args:
n_classes: Number of classes, must be greater than 2 (for 2 classes, use
`binary_classification_head`).
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
label_vocabulary: A list or tuple of strings representing possible label
values. If it is not given, that means labels are already encoded as an
integer within [0, n_classes). If given, labels must be of string type and
have any value in `label_vocabulary`. Note that errors will be raised if
`label_vocabulary` is not provided but labels are strings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how to
reduce training loss over batch. Defaults to `SUM_OVER_BATCH_SIZE`, namely
weighted sum of losses divided by batch size. See `tf.losses.Reduction`.
loss_fn: Optional loss function.
name: name of the head. If provided, summary and metrics keys will be
suffixed by `"/" + name`. Also used as `name_scope` when creating ops.
Returns:
An instance of `_Head` for multi class classification.
Raises:
ValueError: if `n_classes`, `label_vocabulary` or `loss_reduction` is
invalid.
"""
return head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction,
loss_fn=loss_fn,
name=name)
def __init__(self,
model_dir=None,
linear_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
input_layer_partitioner=None,
config=None):
"""Initializes a DNNLinearCombinedClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
linear_feature_columns: An iterable containing all the feature columns
used by linear part of the model. All items in the set must be
instances of classes derived from `FeatureColumn`.
linear_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the linear part of the model. Defaults to FTRL optimizer.
dnn_feature_columns: An iterable containing all the feature columns used
by deep part of the model. All items in the set must be instances of
classes derived from `FeatureColumn`.
dnn_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the deep part of the model. Defaults to Adagrad optimizer.
dnn_hidden_units: List of hidden units per layer. All layers are fully
connected.
dnn_activation_fn: Activation function applied to each layer. If None,
will use `tf.nn.relu`.
dnn_dropout: When not None, the probability we will drop out
a given coordinate.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
input_layer_partitioner: Partitioner for input layer. Defaults to
`min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: RunConfig object to configure the runtime settings.
Raises:
ValueError: If both linear_feature_columns and dnn_features_columns are
empty at the same time.
"""
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = linear_feature_columns + dnn_feature_columns
if not self._feature_columns:
raise ValueError('Either linear_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss() # pylint: disable=protected-access
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes)
def _model_fn(features, labels, mode, config):
return _dnn_linear_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNLinearCombinedClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config)
def __init__(self,
sequence_feature_columns,
context_feature_columns=None,
num_units=None,
cell_type=USE_DEFAULT,
rnn_cell_fn=None,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
loss_reduction=losses.Reduction.SUM_OVER_BATCH_SIZE,
input_layer_partitioner=None,
config=None):
"""Initializes a `RNNClassifier` instance.
Args:
sequence_feature_columns: An iterable containing the `FeatureColumn`s
that represent sequential input. All items in the set should either be
sequence columns (e.g. `sequence_numeric_column`) or constructed from
one (e.g. `embedding_column` with `sequence_categorical_column_*` as
input).
context_feature_columns: An iterable containing the `FeatureColumn`s
for contextual input. The data represented by these columns will be
replicated and given to the RNN at each timestep. These columns must be
instances of classes derived from `_DenseColumn` such as
`numeric_column`, not the sequential variants.
num_units: Iterable of integer number of hidden units per RNN layer. If
set, `cell_type` must also be specified and `rnn_cell_fn` must be
`None`.
cell_type: A subclass of `tf.nn.rnn_cell.RNNCell` or a string specifying
the cell type. Supported strings are: `'basic_rnn'`, `'lstm'`, and
`'gru'`. If set, `num_units` must also be specified and `rnn_cell_fn`
must be `None`.
rnn_cell_fn: A function with one argument, a `tf.estimator.ModeKeys`, and
returns an object of type `tf.nn.rnn_cell.RNNCell` that will be used to
construct the RNN. If set, `num_units` and `cell_type` cannot be set.
This is for advanced users who need additional customization beyond
`num_units` and `cell_type`. Note that `tf.nn.rnn_cell.MultiRNNCell` is
needed for stacked RNNs.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` or string specifying optimizer
type. Defaults to Adagrad optimizer.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM_OVER_BATCH_SIZE`.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
Raises:
ValueError: If `num_units`, `cell_type`, and `rnn_cell_fn` are not
compatible.
"""
rnn_cell_fn = _assert_rnn_cell_fn(rnn_cell_fn, num_units, cell_type)
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
return _rnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
rnn_cell_fn=rnn_cell_fn,
sequence_feature_columns=tuple(sequence_feature_columns or []),
context_feature_columns=tuple(context_feature_columns or []),
optimizer=optimizer,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(RNNClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
model_dir=None,
linear_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None,
warm_start_from=None,
loss_reduction=losses.Reduction.SUM):
"""Initializes a DNNLinearCombinedClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
linear_feature_columns: An iterable containing all the feature columns
used by linear part of the model. All items in the set must be
instances of classes derived from `FeatureColumn`.
linear_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the linear part of the model. Defaults to FTRL optimizer.
dnn_feature_columns: An iterable containing all the feature columns used
by deep part of the model. All items in the set must be instances of
classes derived from `FeatureColumn`.
dnn_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the deep part of the model. Defaults to Adagrad optimizer.
dnn_hidden_units: List of hidden units per layer. All layers are fully
connected.
dnn_activation_fn: Activation function applied to each layer. If None,
will use `tf.nn.relu`.
dnn_dropout: When not None, the probability we will drop out
a given coordinate.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
input_layer_partitioner: Partitioner for input layer. Defaults to
`min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: RunConfig object to configure the runtime settings.
warm_start_from: A string filepath to a checkpoint to warm-start from, or
a `WarmStartSettings` object to fully configure warm-starting. If the
string filepath is provided instead of a `WarmStartSettings`, then all
weights are warm-started, and it is assumed that vocabularies and Tensor
names are unchanged.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM`.
Raises:
ValueError: If both linear_feature_columns and dnn_features_columns are
empty at the same time.
"""
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = (
list(linear_feature_columns) + list(dnn_feature_columns))
if not self._feature_columns:
raise ValueError('Either linear_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
"""Call the _dnn_linear_combined_model_fn."""
return _dnn_linear_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNLinearCombinedClassifier, self).__init__(
model_fn=_model_fn, model_dir=model_dir, config=config,
warm_start_from=warm_start_from)
def __init__(self,
model_dir=None,
linear_feature_columns=None,
linear_optimizer='Ftrl',
dnn_feature_columns=None,
dnn_optimizer='Adagrad',
dnn_hidden_units=None,
dnn_activation_fn=nn.relu,
dnn_dropout=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
input_layer_partitioner=None,
config=None,
warm_start_from=None,
loss_reduction=losses.Reduction.SUM):
"""Initializes a DNNLinearCombinedClassifier instance.
Args:
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
linear_feature_columns: An iterable containing all the feature columns
used by linear part of the model. All items in the set must be
instances of classes derived from `FeatureColumn`.
linear_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the linear part of the model. Defaults to FTRL optimizer.
dnn_feature_columns: An iterable containing all the feature columns used
by deep part of the model. All items in the set must be instances of
classes derived from `FeatureColumn`.
dnn_optimizer: An instance of `tf.Optimizer` used to apply gradients to
the deep part of the model. Defaults to Adagrad optimizer.
dnn_hidden_units: List of hidden units per layer. All layers are fully
connected.
dnn_activation_fn: Activation function applied to each layer. If None,
will use `tf.nn.relu`.
dnn_dropout: When not None, the probability we will drop out
a given coordinate.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
input_layer_partitioner: Partitioner for input layer. Defaults to
`min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: RunConfig object to configure the runtime settings.
warm_start_from: A string filepath to a checkpoint to warm-start from, or
a `WarmStartSettings` object to fully configure warm-starting. If the
string filepath is provided instead of a `WarmStartSettings`, then all
weights are warm-started, and it is assumed that vocabularies and Tensor
names are unchanged.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM`.
Raises:
ValueError: If both linear_feature_columns and dnn_features_columns are
empty at the same time.
"""
linear_feature_columns = linear_feature_columns or []
dnn_feature_columns = dnn_feature_columns or []
self._feature_columns = (list(linear_feature_columns) +
list(dnn_feature_columns))
if not self._feature_columns:
raise ValueError(
'Either linear_feature_columns or dnn_feature_columns '
'must be defined.')
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
"""Call the _dnn_linear_combined_model_fn."""
return _dnn_linear_combined_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
linear_feature_columns=linear_feature_columns,
linear_optimizer=linear_optimizer,
dnn_feature_columns=dnn_feature_columns,
dnn_optimizer=dnn_optimizer,
dnn_hidden_units=dnn_hidden_units,
dnn_activation_fn=dnn_activation_fn,
dnn_dropout=dnn_dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNLinearCombinedClassifier,
self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config,
warm_start_from=warm_start_from)
def __init__(self,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Ftrl',
config=None,
partitioner=None):
"""Construct a `LinearClassifier` estimator object.
Args:
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
Note that class labels are integers representing the class index (i.e.
values from 0 to n_classes-1). For arbitrary label values (e.g. string
labels), convert to class indices first.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Defaults
to FTRL optimizer.
config: `RunConfig` object to configure the runtime settings.
partitioner: Optional. Partitioner for input layer.
Returns:
A `LinearClassifier` estimator.
Raises:
ValueError: if n_classes < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _model_fn(features, labels, mode, config):
return _linear_model_fn(features=features,
labels=labels,
mode=mode,
head=head,
feature_columns=tuple(feature_columns
or []),
optimizer=optimizer,
partitioner=partitioner,
config=config)
super(LinearClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
hidden_units,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Adagrad',
activation_fn=nn.relu,
dropout=None,
input_layer_partitioner=None,
config=None):
"""Initializes a `DNNClassifier` instance.
Args:
hidden_units: Iterable of number hidden units per layer. All layers are
fully connected. Ex. `[64, 32]` means first layer has 64 nodes and
second one has 32.
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `_FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator to
continue training a previously saved model.
n_classes: Number of label classes. Defaults to 2, namely binary
classification. Must be > 1.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Defaults
to Adagrad optimizer.
activation_fn: Activation function applied to each layer. If `None`, will
use `tf.nn.relu`.
dropout: When not `None`, the probability we will drop out a given
coordinate.
input_layer_partitioner: Optional. Partitioner for input layer. Defaults
to `min_max_variable_partitioner` with `min_slice_size` 64 << 20.
config: `RunConfig` object to configure the runtime settings.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary)
def _model_fn(features, labels, mode, config):
return _dnn_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
hidden_units=hidden_units,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
activation_fn=activation_fn,
dropout=dropout,
input_layer_partitioner=input_layer_partitioner,
config=config)
super(DNNClassifier, self).__init__(model_fn=_model_fn,
model_dir=model_dir,
config=config)
def __init__(self,
feature_columns,
model_dir=None,
n_classes=2,
weight_column=None,
label_vocabulary=None,
optimizer='Ftrl',
config=None,
partitioner=None,
warm_start_from=None,
loss_reduction=losses.Reduction.SUM,
sparse_combiner='sum'):
"""Construct a `LinearClassifier` estimator object.
Args:
feature_columns: An iterable containing all the feature columns used by
the model. All items in the set should be instances of classes derived
from `FeatureColumn`.
model_dir: Directory to save model parameters, graph and etc. This can
also be used to load checkpoints from the directory into a estimator
to continue training a previously saved model.
n_classes: number of label classes. Default is binary classification.
Note that class labels are integers representing the class index (i.e.
values from 0 to n_classes-1). For arbitrary label values (e.g. string
labels), convert to class indices first.
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example. If it is a string, it is
used as a key to fetch weight tensor from the `features`. If it is a
`_NumericColumn`, raw tensor is fetched by key `weight_column.key`,
then weight_column.normalizer_fn is applied on it to get weight tensor.
label_vocabulary: A list of strings represents possible label values. If
given, labels must be string type and have any value in
`label_vocabulary`. If it is not given, that means labels are
already encoded as integer or float within [0, 1] for `n_classes=2` and
encoded as integer values in {0, 1,..., n_classes-1} for `n_classes`>2 .
Also there will be errors if vocabulary is not provided and labels are
string.
optimizer: An instance of `tf.Optimizer` used to train the model. Can also
be a string (one of 'Adagrad', 'Adam', 'Ftrl', 'RMSProp', 'SGD'), or
callable. Defaults to FTRL optimizer.
config: `RunConfig` object to configure the runtime settings.
partitioner: Optional. Partitioner for input layer.
warm_start_from: A string filepath to a checkpoint to warm-start from, or
a `WarmStartSettings` object to fully configure warm-starting. If the
string filepath is provided instead of a `WarmStartSettings`, then all
weights and biases are warm-started, and it is assumed that vocabularies
and Tensor names are unchanged.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how
to reduce training loss over batch. Defaults to `SUM`.
sparse_combiner: A string specifying how to reduce if a categorical column
is multivalent. One of "mean", "sqrtn", and "sum" -- these are
effectively different ways to do example-level normalization, which can
be useful for bag-of-words features. for more details, see
`tf.feature_column.linear_model`.
Returns:
A `LinearClassifier` estimator.
Raises:
ValueError: if n_classes < 2.
"""
if n_classes == 2:
head = head_lib._binary_logistic_head_with_sigmoid_cross_entropy_loss( # pylint: disable=protected-access
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
else:
head = head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint: disable=protected-access
n_classes, weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction)
def _model_fn(features, labels, mode, config):
"""Call the defined shared _linear_model_fn."""
return _linear_model_fn(
features=features,
labels=labels,
mode=mode,
head=head,
feature_columns=tuple(feature_columns or []),
optimizer=optimizer,
partitioner=partitioner,
config=config,
sparse_combiner=sparse_combiner)
super(LinearClassifier, self).__init__(
model_fn=_model_fn,
model_dir=model_dir,
config=config,
warm_start_from=warm_start_from)
def multi_class_head(n_classes,
weight_column=None,
label_vocabulary=None,
loss_reduction=losses.Reduction.SUM_OVER_BATCH_SIZE,
loss_fn=None,
name=None):
"""Creates a `_Head` for multi class classification.
Uses `sparse_softmax_cross_entropy` loss.
The head expects `logits` with shape `[D0, D1, ... DN, n_classes]`.
In many applications, the shape is `[batch_size, n_classes]`.
`labels` must be a dense `Tensor` with shape matching `logits`, namely
`[D0, D1, ... DN, 1]`. If `label_vocabulary` given, `labels` must be a string
`Tensor` with values from the vocabulary. If `label_vocabulary` is not given,
`labels` must be an integer `Tensor` with values specifying the class index.
If `weight_column` is specified, weights must be of shape
`[D0, D1, ... DN]`, or `[D0, D1, ... DN, 1]`.
The loss is the weighted sum over the input dimensions. Namely, if the input
labels have shape `[batch_size, 1]`, the loss is the weighted sum over
`batch_size`.
Also supports custom `loss_fn`. `loss_fn` takes `(labels, logits)` or
`(labels, logits, features)` as arguments and returns unreduced loss with
shape `[D0, D1, ... DN, 1]`. `loss_fn` must support integer `labels` with
shape `[D0, D1, ... DN, 1]`. Namely, the head applies `label_vocabulary` to
the input labels before passing them to `loss_fn`.
The head can be used with a canned estimator. Example:
```python
my_head = tf.contrib.estimator.multi_class_head(n_classes=3)
my_estimator = tf.contrib.estimator.DNNEstimator(
head=my_head,
hidden_units=...,
feature_columns=...)
```
It can also be used with a custom `model_fn`. Example:
```python
def _my_model_fn(features, labels, mode):
my_head = tf.contrib.estimator.multi_class_head(n_classes=3)
logits = tf.keras.Model(...)(features)
return my_head.create_estimator_spec(
features=features,
mode=mode,
labels=labels,
optimizer=tf.AdagradOptimizer(learning_rate=0.1),
logits=logits)
my_estimator = tf.estimator.Estimator(model_fn=_my_model_fn)
```
Args:
n_classes: Number of classes, must be greater than 2 (for 2 classes, use
`binary_classification_head`).
weight_column: A string or a `_NumericColumn` created by
`tf.feature_column.numeric_column` defining feature column representing
weights. It is used to down weight or boost examples during training. It
will be multiplied by the loss of the example.
label_vocabulary: A list or tuple of strings representing possible label
values. If it is not given, that means labels are already encoded as an
integer within [0, n_classes). If given, labels must be of string type and
have any value in `label_vocabulary`. Note that errors will be raised if
`label_vocabulary` is not provided but labels are strings.
loss_reduction: One of `tf.losses.Reduction` except `NONE`. Describes how to
reduce training loss over batch. Defaults to `SUM_OVER_BATCH_SIZE`, namely
weighted sum of losses divided by batch size. See `tf.losses.Reduction`.
loss_fn: Optional loss function.
name: name of the head. If provided, summary and metrics keys will be
suffixed by `"/" + name`. Also used as `name_scope` when creating ops.
Returns:
An instance of `_Head` for multi class classification.
Raises:
ValueError: if `n_classes`, `label_vocabulary` or `loss_reduction` is
invalid.
"""
return head_lib._multi_class_head_with_softmax_cross_entropy_loss( # pylint:disable=protected-access
n_classes=n_classes,
weight_column=weight_column,
label_vocabulary=label_vocabulary,
loss_reduction=loss_reduction,
loss_fn=loss_fn,
name=name)
