TF 2.0 is out! Get hands-on practice at TF World, Oct 28-31. Use code TF20 for 20% off select passes. Register now

tfp.optimizer.VariationalSGD

View source on GitHub

Class VariationalSGD

An optimizer module for constant stochastic gradient descent.

This implements an optimizer module for the constant stochastic gradient descent algorithm [(Mandt et al., 2017)][1]. The optimization variable is regarded as an approximate sample from the posterior .

Args:

  • batch_size: Scalar int-like Tensor. The number of examples in a minibatch in the data set. Note: Assumes the loss is taken as the mean over a minibatch. Otherwise if the sum was taken set this to 1.
  • total_num_examples: Scalar int-like Tensor. The total number of examples in the data set.
  • max_learning_rate: Scalar float-like Tensor. A maximum allowable effective coordinate-wise learning rate. The algorithm scales down any effective learning rate (i.e. after preconditioning) that is larger than this. (Default: 1)
  • preconditioner_decay_rate: Scalar float-like Tensor. The exponential decay rate of the rescaling of the preconditioner (RMSprop). (This is "alpha" in Mandt et al. (2017)). Should be smaller than but nearly 1 to approximate sampling from the posterior. (Default: 0.95)
  • burnin: Scalar int-like Tensor. The number of iterations to collect gradient statistics to update the preconditioner before starting to draw noisy samples. (Default: 25)
  • burnin_max_learning_rate: Scalar float-like Tensor. Maximum learning rate to use during the burnin period. (Default: 1e-8)
  • use_single_learning_rate: Boolean Indicates whether one single learning rate is used or coordinate_wise learning rates are used. (Default: False)
  • name: Python str describing ops managed by this function. (Default: "VariationalSGD")

Raises:

  • InvalidArgumentError: If preconditioner_decay_rate is a Tensor not in (0,1].

References

[1]: Stephan Mandt, Matthew D. Hoffman, and David M. Blei. Stochastic Gradient Descent as Approximate Bayesian Inference. arXiv preprint arXiv:1704.04289, 2017. https://arxiv.org/abs/1704.04289

__init__

View source

__init__(
    batch_size,
    total_num_examples,
    max_learning_rate=1.0,
    preconditioner_decay_rate=0.95,
    burnin=25,
    burnin_max_learning_rate=1e-06,
    use_single_learning_rate=False,
    name=None
)

Create a new Optimizer.

This must be called by the constructors of subclasses. Note that Optimizer instances should not bind to a single graph, and so shouldn't keep Tensors as member variables. Generally you should be able to use the _set_hyper()/state.get_hyper() facility instead.

This class in stateful and thread-compatible.

Args:

  • name: A non-empty string. The name to use for accumulators created for the optimizer.
  • **kwargs: keyword arguments. Allowed to be {clipnorm, clipvalue, lr, decay}. clipnorm is clip gradients by norm; clipvalue is clip gradients by value, decay is included for backward compatibility to allow time inverse decay of learning rate. lr is included for backward compatibility, recommended to use learning_rate instead.

Raises:

  • ValueError: If name is malformed.
  • RuntimeError: If _create_slots has been overridden instead of _create_vars.

Properties

iterations

Variable. The number of training steps this Optimizer has run.

weights

Returns variables of this Optimizer based on the order created.

Methods

add_slot

add_slot(
    var,
    slot_name,
    initializer='zeros'
)

Add a new slot variable for var.

add_weight

add_weight(
    name,
    shape,
    dtype=None,
    initializer='zeros',
    trainable=None,
    synchronization=tf_variables.VariableSynchronization.AUTO,
    aggregation=tf_variables.VariableAggregation.NONE
)

apply_gradients

apply_gradients(
    grads_and_vars,
    name=None
)

Apply gradients to variables.

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

Args:

  • grads_and_vars: List of (gradient, variable) pairs.
  • name: Optional name for the returned operation. Default to the name passed to the Optimizer constructor.

Returns:

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

Raises:

  • TypeError: If grads_and_vars is malformed.
  • ValueError: If none of the variables have gradients.

from_config

from_config(
    cls,
    config,
    custom_objects=None
)

Creates an optimizer from its config.

This method is the reverse of get_config, capable of instantiating the same optimizer from the config dictionary.

Arguments:

  • config: A Python dictionary, typically the output of get_config.
  • custom_objects: A Python dictionary mapping names to additional Python objects used to create this optimizer, such as a function used for a hyperparameter.

Returns:

An optimizer instance.

get_config

View source

get_config()

Returns the config of the optimimizer.

An optimizer config is a Python dictionary (serializable) containing the configuration of an optimizer. The same optimizer can be reinstantiated later (without any saved state) from this configuration.

Returns:

Python dictionary.

get_gradients

get_gradients(
    loss,
    params
)

Returns gradients of loss with respect to params.

Arguments:

  • loss: Loss tensor.
  • params: List of variables.

Returns:

List of gradient tensors.

Raises:

  • ValueError: In case any gradient cannot be computed (e.g. if gradient function not implemented).

get_slot

get_slot(
    var,
    slot_name
)

get_slot_names

get_slot_names()

A list of names for this optimizer's slots.

get_updates

get_updates(
    loss,
    params
)

get_weights

get_weights()

minimize

minimize(
    loss,
    var_list,
    grad_loss=None,
    name=None
)

Minimize loss by updating var_list.

This method simply computes gradient using tf.GradientTape and calls apply_gradients(). If you want to process the gradient before applying then call tf.GradientTape and apply_gradients() explicitly instead of using this function.

Args:

  • loss: A callable taking no arguments which returns the value to minimize.
  • var_list: list or tuple of Variable objects to update to minimize loss, or a callable returning the list or tuple of Variable objects. Use callable when the variable list would otherwise be incomplete before minimize since the variables are created at the first time loss is called.
  • grad_loss: Optional. A Tensor holding the gradient computed for loss.
  • name: Optional name for the returned operation.

Returns:

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

Raises:

  • ValueError: If some of the variables are not Variable objects.

set_weights

set_weights(weights)

variables

variables()

Returns variables of this Optimizer based on the order created.