tf.contrib.kfac.fisher_factors.FullFactor

Class FullFactor

Defined in tensorflow/contrib/kfac/python/ops/fisher_factors.py.

FisherFactor for a full matrix representation of the Fisher of a parameter.

Note that this uses the naive "square the sum estimator", and so is applicable to any type of parameter in principle, but has very high variance.

Properties

name

Methods

__init__

__init__(
    params_grads,
    batch_size
)

get_cholesky

get_cholesky(damping_func)

get_cholesky_inverse

get_cholesky_inverse(damping_func)

get_cov

get_cov()

get_cov_as_linear_operator

get_cov_as_linear_operator()

get_eigendecomp

get_eigendecomp()

Creates or retrieves eigendecomposition of self._cov.

get_inverse

get_inverse(damping_func)

get_matpower

get_matpower(
    exp,
    damping_func
)

instantiate_cov_variables

instantiate_cov_variables()

Makes the internal cov variable(s).

instantiate_inv_variables

instantiate_inv_variables()

Makes the internal "inverse" variable(s).

make_covariance_update_op

make_covariance_update_op(ema_decay)

Constructs and returns the covariance update Op.

Args:

  • ema_decay: The exponential moving average decay (float or Tensor).

Returns:

An Op for updating the covariance Variable referenced by _cov.

make_inverse_update_ops

make_inverse_update_ops()

Create and return update ops corresponding to registered computations.

register_cholesky

register_cholesky(damping_func)

Registers a Cholesky factor to be maintained and served on demand.

This creates a variable and signals make_inverse_update_ops to make the corresponding update op. The variable can be read via the method get_cholesky.

Args:

  • damping_func: A function that computes a 0-D Tensor or a float which will be the damping value used. i.e. damping = damping_func().

register_cholesky_inverse

register_cholesky_inverse(damping_func)

Registers an inverse Cholesky factor to be maintained/served on demand.

This creates a variable and signals make_inverse_update_ops to make the corresponding update op. The variable can be read via the method get_cholesky_inverse.

Args:

  • damping_func: A function that computes a 0-D Tensor or a float which will be the damping value used. i.e. damping = damping_func().

register_inverse

register_inverse(damping_func)

register_matpower

register_matpower(
    exp,
    damping_func
)

Registers a matrix power to be maintained and served on demand.

This creates a variable and signals make_inverse_update_ops to make the corresponding update op. The variable can be read via the method get_matpower.

Args:

  • exp: float. The exponent to use in the matrix power.
  • damping_func: A function that computes a 0-D Tensor or a float which will be the damping value used. i.e. damping = damping_func().