tf.compat.v1.train.FtrlOptimizer

Optimizer that implements the FTRL algorithm.

Inherits From: Optimizer

tf.compat.v1.train.FtrlOptimizer(
    learning_rate,
    learning_rate_power=-0.5,
    initial_accumulator_value=0.1,
    l1_regularization_strength=0.0,
    l2_regularization_strength=0.0,
    use_locking=False,
    name='Ftrl',
    accum_name=None,
    linear_name=None,
    l2_shrinkage_regularization_strength=0.0,
    beta=None
)

This version has support for both online L2 (McMahan et al., 2013) and shrinkage-type L2, which is the addition of an L2 penalty to the loss function.

Ad-click prediction: McMahan et al., 2013 (pdf)

learning_rateA float value or a constant float Tensor.
learning_rate_powerA float value, must be less or equal to zero. Controls how the learning rate decreases during training. Use zero for a fixed learning rate. See section 3.1 in (McMahan et al., 2013).
initial_accumulator_valueThe starting value for accumulators. Only zero or positive values are allowed.
l1_regularization_strengthA float value, must be greater than or equal to zero.
l2_regularization_strengthA float value, must be greater than or equal to zero.
use_lockingIf True use locks for update operations.
nameOptional name prefix for the operations created when applying gradients. Defaults to "Ftrl".
accum_nameThe suffix for the variable that keeps the gradient squared accumulator. If not present, defaults to name.
linear_nameThe suffix for the variable that keeps the linear gradient accumulator. If not present, defaults to name + "_1".
l2_shrinkage_regularization_strengthA float value, must be greater than or equal to zero. This differs from L2 above in that the L2 above is a stabilization penalty, whereas this L2 shrinkage is a magnitude penalty. The FTRL formulation can be written as: w_{t+1} = argminw(\hat{g}{1:t}w + L1||w||_1 + L2||w||_2^2), where \hat{g} = g + (2L2_shrinkagew), and g is the gradient of the loss function w.r.t. the weights w. Specifically, in the absence of L1 regularization, it is equivalent to the following update rule: w_{t+1} = w_t - lr_t / (beta + 2L2lr_t) * g_t - 2L2_shrinkagelr_t / (beta + 2L2lr_t) * w_t where lr_t is the learning rate at t. When input is sparse shrinkage will only happen on the active weights.
betaA float value; corresponds to the beta parameter in the paper.

ValueErrorIf one of the arguments is invalid.

Methods

apply_gradients

View source

apply_gradients(
    grads_and_vars,
    global_step=None,
    name=None,
    skip_gradients_aggregation=False
)

Apply gradients to variables.

This is the second part of minimize(). It returns an Operation that applies gradients.

@compatibility(TF2)

How to Map Arguments

TF1 Arg NameTF2 Arg NameNote
grads_and_varsgrads_and_vars-
global_stepNot supported.Use optimizer.iterations
namename.-

Args
grads_and_varsList of (gradient, variable) pairs as returned by compute_gradients().
global_stepOptional Variable to increment by one after the variables have been updated.
nameOptional name for the returned operation. Default to the name passed to the Optimizer constructor.
skip_gradients_aggregationIf true, gradients aggregation will not be performed inside optimizer. Usually this arg is set to True when you write custom code aggregating gradients outside the optimizer.

Returns
An Operation that applies the specified gradients. If global_step was not None, that operation also increments global_step.

Raises
TypeErrorIf grads_and_vars is malformed.
ValueErrorIf none of the variables have gradients.
RuntimeErrorIf you should use _distributed_apply() instead.

compute_gradients

View source

compute_gradients(
    loss,
    var_list=None,
    gate_gradients=GATE_OP,
    aggregation_method=None,
    colocate_gradients_with_ops=False,
    grad_loss=None
)

Compute gradients of loss for the variables in var_list.

Migrate to TF2

tf.keras.optimizers.Optimizer in TF2 does not provide a compute_gradients method, and you should use a tf.GradientTape to obtain the gradients:

@tf.function
def train step(inputs):
  batch_data, labels = inputs
  with tf.GradientTape() as tape:
    predictions = model(batch_data, training=True)
    loss = tf.keras.losses.CategoricalCrossentropy(
        reduction=tf.keras.losses.Reduction.NONE)(labels, predictions)
  gradients = tape.gradient(loss, model.trainable_variables)
  optimizer.apply_gradients(zip(gradients, model.trainable_variables))

Args: loss: A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable. var_list: Optional list or tuple of tf.Variable to update to minimize loss. Defaults to the list of variables collected in the graph under the key GraphKeys.TRAINABLE_VARIABLES. gate_gradients: How to gate the computation of gradients. Can be GATE_NONE, GATE_OP, or GATE_GRAPH. aggregation_method: Specifies the method used to combine gradient terms. Valid values are defined in the class AggregationMethod. colocate_gradients_with_ops: If True, try colocating gradients with the corresponding op. grad_loss: Optional. A Tensor holding the gradient computed for loss.

Returns: A list of (gradient, variable) pairs. Variable is always present, but gradient can be None.

Raises: TypeError: If var_list contains anything else than Variable objects. ValueError: If some arguments are invalid. RuntimeError: If called with eager execution enabled and loss is not callable.

@compatibility(eager) When eager execution is enabled, gate_gradients, aggregation_method, and colocate_gradients_with_ops are ignored.

Description

This is the first part of minimize(). It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a Tensor, an IndexedSlices, or None if there is no gradient for the given variable.

get_name

View source

get_name()

get_slot

View source

get_slot(
    var, name
)

Return a slot named name created for var by the Optimizer.

Some Optimizer subclasses use additional variables. For example Momentum and Adagrad use variables to accumulate updates. This method gives access to these Variable objects if for some reason you need them.

Use get_slot_names() to get the list of slot names created by the Optimizer.

Args
varA variable passed to minimize() or apply_gradients().
nameA string.

Returns
The Variable for the slot if it was created, None otherwise.

get_slot_names

View source

get_slot_names()

Return a list of the names of slots created by the Optimizer.

See get_slot().

Returns
A list of strings.

minimize

View source

minimize(
    loss,
    global_step=None,
    var_list=None,
    gate_gradients=GATE_OP,
    aggregation_method=None,
    colocate_gradients_with_ops=False,
    name=None,
    grad_loss=None
)

Add operations to minimize loss by updating var_list.

This method simply combines calls compute_gradients() and apply_gradients(). If you want to process the gradient before applying them call compute_gradients() and apply_gradients() explicitly instead of using this function.

Args
lossA Tensor containing the value to minimize.
global_stepOptional Variable to increment by one after the variables have been updated.
var_listOptional list or tuple of Variable objects to update to minimize loss. Defaults to the list of variables collected in the graph under the key GraphKeys.TRAINABLE_VARIABLES.
gate_gradientsHow to gate the computation of gradients. Can be GATE_NONE, GATE_OP, or GATE_GRAPH.
aggregation_methodSpecifies the method used to combine gradient terms. Valid values are defined in the class AggregationMethod.
colocate_gradients_with_opsIf True, try colocating gradients with the corresponding op.
nameOptional name for the returned operation.
grad_lossOptional. A Tensor holding the gradient computed for loss.

Returns
An Operation that updates the variables in var_list. If global_step was not None, that operation also increments global_step.

Raises
ValueErrorIf some of the variables are not Variable objects.

eager compatibility

When eager execution is enabled, loss should be a Python function that takes no arguments and computes the value to be minimized. Minimization (and gradient computation) is done with respect to the elements of var_list if not None, else with respect to any trainable variables created during the execution of the loss function. gate_gradients, aggregation_method, colocate_gradients_with_ops and grad_loss are ignored when eager execution is enabled.

variables

View source

variables()

A list of variables which encode the current state of Optimizer.

Includes slot variables and additional global variables created by the optimizer in the current default graph.

Returns
A list of variables.

GATE_GRAPH2
GATE_NONE0
GATE_OP1