Class DNNEstimator
Inherits From: Estimator
Defined in tensorflow/contrib/learn/python/learn/estimators/dnn.py
.
A Estimator for TensorFlow DNN models with user specified _Head.
THIS CLASS IS DEPRECATED. See contrib/learn/README.md for general migration instructions.
Example:
sparse_feature_a = sparse_column_with_hash_bucket(...)
sparse_feature_b = sparse_column_with_hash_bucket(...)
sparse_feature_a_emb = embedding_column(sparse_id_column=sparse_feature_a,
...)
sparse_feature_b_emb = embedding_column(sparse_id_column=sparse_feature_b,
...)
To create a DNNEstimator for binary classification, where
estimator = DNNEstimator(
feature_columns=[sparse_feature_a_emb, sparse_feature_b_emb],
head=tf.contrib.learn.multi_class_head(n_classes=2),
hidden_units=[1024, 512, 256])
If your label is keyed with "y" in your labels dict, and weights are keyed
with "w" in features dict, and you want to enable centered bias,
head = tf.contrib.learn.multi_class_head(
n_classes=2,
label_name="x",
weight_column_name="w",
enable_centered_bias=True)
estimator = DNNEstimator(
feature_columns=[sparse_feature_a_emb, sparse_feature_b_emb],
head=head,
hidden_units=[1024, 512, 256])
# Input builders
def input_fn_train: # returns x, y (where y represents label's class index).
pass
estimator.fit(input_fn=input_fn_train)
def input_fn_eval: # returns x, y (where y represents label's class index).
pass
estimator.evaluate(input_fn=input_fn_eval)
estimator.predict(x=x) # returns predicted labels (i.e. label's class index).
Input of fit
and evaluate
should have following features,
otherwise there will be a KeyError
:
- if
weight_column_name
is notNone
, a feature withkey=weight_column_name
whose value is aTensor
. - for each
column
infeature_columns
:- if
column
is aSparseColumn
, a feature withkey=column.name
whosevalue
is aSparseTensor
. - if
column
is aWeightedSparseColumn
, two features: the first withkey
the id column name, the second withkey
the weight column name. Both features'value
must be aSparseTensor
. - if
column
is aRealValuedColumn
, a feature withkey=column.name
whosevalue
is aTensor
.
- if
__init__
__init__(
head,
hidden_units,
feature_columns,
model_dir=None,
optimizer=None,
activation_fn=tf.nn.relu,
dropout=None,
gradient_clip_norm=None,
config=None,
feature_engineering_fn=None,
embedding_lr_multipliers=None,
input_layer_min_slice_size=None
)
Initializes a DNNEstimator
instance.
Args:
head
:Head
instance.hidden_units
: List of 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 fromFeatureColumn
.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.optimizer
: An instance oftf.Optimizer
used to train the model. IfNone
, will use an Adagrad optimizer.activation_fn
: Activation function applied to each layer. IfNone
, will usetf.nn.relu
. Note that a string containing the unqualified name of the op may also be provided, e.g., "relu", "tanh", or "sigmoid".dropout
: When notNone
, the probability we will drop out a given coordinate.gradient_clip_norm
: A float > 0. If provided, gradients are clipped to their global norm with this clipping ratio. Seetf.clip_by_global_norm
for more details.config
:RunConfig
object to configure the runtime settings.feature_engineering_fn
: Feature engineering function. Takes features and labels which are the output ofinput_fn
and returns features and labels which will be fed into the model.embedding_lr_multipliers
: Optional. A dictionary fromEmbeddingColumn
to afloat
multiplier. Multiplier will be used to multiply with learning rate for the embedding variables.input_layer_min_slice_size
: Optional. The min slice size of input layer partitions. If not provided, will use the default of 64M.
Returns:
A DNNEstimator
estimator.
Properties
config
model_dir
Returns a path in which the eval process will look for checkpoints.
model_fn
Returns the model_fn which is bound to self.params.
Returns:
The model_fn with the following signature:
def model_fn(features, labels, mode, metrics)
Methods
tf.contrib.learn.DNNEstimator.evaluate
evaluate(
x=None,
y=None,
input_fn=None,
feed_fn=None,
batch_size=None,
steps=None,
metrics=None,
name=None,
checkpoint_path=None,
hooks=None,
log_progress=True
)
See Evaluable
. (deprecated arguments)
Raises:
ValueError
: If at least one ofx
ory
is provided, and at least one ofinput_fn
orfeed_fn
is provided. Or ifmetrics
is notNone
ordict
.
tf.contrib.learn.DNNEstimator.export
export(
export_dir,
input_fn=export._default_input_fn,
input_feature_key=None,
use_deprecated_input_fn=True,
signature_fn=None,
prediction_key=None,
default_batch_size=1,
exports_to_keep=None,
checkpoint_path=None
)
Exports inference graph into given dir. (deprecated)
Args:
export_dir
: A string containing a directory to write the exported graph and checkpoints.input_fn
: Ifuse_deprecated_input_fn
is true, then a function that givenTensor
ofExample
strings, parses it into features that are then passed to the model. Otherwise, a function that takes no argument and returns a tuple of (features, labels), where features is a dict of string key toTensor
and labels is aTensor
that's currently not used (and so can beNone
).input_feature_key
: Only used ifuse_deprecated_input_fn
is false. String key into the features dict returned byinput_fn
that corresponds to a the rawExample
stringsTensor
that the exported model will take as input. Can only beNone
if you're using a customsignature_fn
that does not use the first arg (examples).use_deprecated_input_fn
: Determines the signature format ofinput_fn
.signature_fn
: Function that returns a default signature and a named signature map, givenTensor
ofExample
strings,dict
ofTensor
s for features andTensor
ordict
ofTensor
s for predictions.prediction_key
: The key for a tensor in thepredictions
dict (output from themodel_fn
) to use as thepredictions
input to thesignature_fn
. Optional. IfNone
, predictions will pass tosignature_fn
without filtering.default_batch_size
: Default batch size of theExample
placeholder.exports_to_keep
: Number of exports to keep.checkpoint_path
: the checkpoint path of the model to be exported. If it isNone
(which is default), will use the latest checkpoint in export_dir.
Returns:
The string path to the exported directory. NB: this functionality was added ca. 2016/09/25; clients that depend on the return value may need to handle the case where this function returns None because subclasses are not returning a value.
tf.contrib.learn.DNNEstimator.export_savedmodel
export_savedmodel(
export_dir_base,
serving_input_fn,
default_output_alternative_key=None,
assets_extra=None,
as_text=False,
checkpoint_path=None,
graph_rewrite_specs=(GraphRewriteSpec((tag_constants.SERVING,), ()),),
strip_default_attrs=False
)
Exports inference graph as a SavedModel into given dir.
Args:
export_dir_base
: A string containing a directory to write the exported graph and checkpoints.serving_input_fn
: A function that takes no argument and returns anInputFnOps
.default_output_alternative_key
: the name of the head to serve when none is specified. Not needed for single-headed models.assets_extra
: A dict specifying how to populate the assets.extra directory within the exported SavedModel. Each key should give the destination path (including the filename) relative to the assets.extra directory. The corresponding value gives the full path of the source file to be copied. For example, the simple case of copying a single file without renaming it is specified as{'my_asset_file.txt': '/path/to/my_asset_file.txt'}
.as_text
: whether to write the SavedModel proto in text format.checkpoint_path
: The checkpoint path to export. If None (the default), the most recent checkpoint found within the model directory is chosen.graph_rewrite_specs
: an iterable ofGraphRewriteSpec
. Each element will produce a separate MetaGraphDef within the exported SavedModel, tagged and rewritten as specified. Defaults to a single entry using the default serving tag ("serve") and no rewriting.strip_default_attrs
: Boolean. IfTrue
, default-valued attributes will be removed from the NodeDefs. For a detailed guide, see Stripping Default-Valued Attributes.
Returns:
The string path to the exported directory.
Raises:
ValueError
: if an unrecognized export_type is requested.
tf.contrib.learn.DNNEstimator.fit
fit(
x=None,
y=None,
input_fn=None,
steps=None,
batch_size=None,
monitors=None,
max_steps=None
)
See Trainable
. (deprecated arguments)
Raises:
ValueError
: Ifx
ory
are notNone
whileinput_fn
is notNone
.ValueError
: If bothsteps
andmax_steps
are notNone
.
tf.contrib.learn.DNNEstimator.get_params
get_params(deep=True)
Get parameters for this estimator.
Args:
deep
: boolean, optionalIf
True
, will return the parameters for this estimator and contained subobjects that are estimators.
Returns:
params
: mapping of string to any Parameter names mapped to their values.
tf.contrib.learn.DNNEstimator.get_variable_names
get_variable_names()
Returns list of all variable names in this model.
Returns:
List of names.
tf.contrib.learn.DNNEstimator.get_variable_value
get_variable_value(name)
Returns value of the variable given by name.
Args:
name
: string, name of the tensor.
Returns:
Numpy array - value of the tensor.
tf.contrib.learn.DNNEstimator.partial_fit
partial_fit(
x=None,
y=None,
input_fn=None,
steps=1,
batch_size=None,
monitors=None
)
Incremental fit on a batch of samples. (deprecated arguments)
This method is expected to be called several times consecutively on different or the same chunks of the dataset. This either can implement iterative training or out-of-core/online training.
This is especially useful when the whole dataset is too big to fit in memory at the same time. Or when model is taking long time to converge, and you want to split up training into subparts.
Args:
x
: Matrix of shape [n_samples, n_features...]. Can be iterator that returns arrays of features. The training input samples for fitting the model. If set,input_fn
must beNone
.y
: Vector or matrix [n_samples] or [n_samples, n_outputs]. Can be iterator that returns array of labels. The training label values (class labels in classification, real numbers in regression). If set,input_fn
must beNone
.input_fn
: Input function. If set,x
,y
, andbatch_size
must beNone
.steps
: Number of steps for which to train model. IfNone
, train forever.batch_size
: minibatch size to use on the input, defaults to first dimension ofx
. Must beNone
ifinput_fn
is provided.monitors
: List ofBaseMonitor
subclass instances. Used for callbacks inside the training loop.
Returns:
self
, for chaining.
Raises:
ValueError
: If at least one ofx
andy
is provided, andinput_fn
is provided.
tf.contrib.learn.DNNEstimator.predict
predict(
x=None,
input_fn=None,
batch_size=None,
outputs=None,
as_iterable=True,
iterate_batches=False
)
Returns predictions for given features. (deprecated arguments)
Args:
x
: Matrix of shape [n_samples, n_features...]. Can be iterator that returns arrays of features. The training input samples for fitting the model. If set,input_fn
must beNone
.input_fn
: Input function. If set,x
and 'batch_size' must beNone
.batch_size
: Override default batch size. If set, 'input_fn' must be 'None'.outputs
: list ofstr
, name of the output to predict. IfNone
, returns all.as_iterable
: If True, return an iterable which keeps yielding predictions for each example until inputs are exhausted. Note: The inputs must terminate if you want the iterable to terminate (e.g. be sure to pass num_epochs=1 if you are using something like read_batch_features).iterate_batches
: If True, yield the whole batch at once instead of decomposing the batch into individual samples. Only relevant when as_iterable is True.
Returns:
A numpy array of predicted classes or regression values if the
constructor's model_fn
returns a Tensor
for predictions
or a dict
of numpy arrays if model_fn
returns a dict
. Returns an iterable of
predictions if as_iterable is True.
Raises:
ValueError
: If x and input_fn are both provided or bothNone
.
tf.contrib.learn.DNNEstimator.set_params
set_params(**params)
Set the parameters of this estimator.
The method works on simple estimators as well as on nested objects
(such as pipelines). The former have parameters of the form
<component>__<parameter>
so that it's possible to update each
component of a nested object.
Args:
**params
: Parameters.
Returns:
self
Raises:
ValueError
: If params contain invalid names.