tfConstrainedGauss.solve_id package

Submodules

tfConstrainedGauss.solve_id.model_id module

class tfConstrainedGauss.solve_id.model_id.LayerMultPrecCov(*args, **kwargs)

Bases: tensorflow.python.keras.engine.base_layer.Layer

Layer that multiplies precision by covarinace matrix

__init__(n: int, non_zero_idx_pairs: List[Tuple[int, int]], non_zero_vals: numpy.array, **kwargs)

Constructor

Args:

n (int): Size of matrix non_zero_idx_pairs (List[Tuple[int,int]]): List of non-zero index pairs in the matrix non_zero_vals (np.array): Non-zero values in the precision matrix corresponding to non_zero_idx_pairs

call(inputs)

Call the layer

Args:

inputs ([type]): Covariance matrix: batch_size x n x n

Returns:

[type]: Product of precision and covariance matrix: vector of length batch_size

classmethod constructDiag(n: int, non_zero_idx_pairs: List[Tuple[int, int]], init_diag_val: float = 1.0, **kwargs)

Constructor for a diagonal precision matrix with elements specified

Args:

n (int): Size of matrix non_zero_idx_pairs (List[Tuple[int,int]]): List of non-zero index pairs in the matrix init_diag_val (float, optional): Value of the diagonal part of precision matrix. Defaults to 1.0.

Returns:

LayerMultPrecCov: layer

classmethod from_config(config)

Construct from config

get_config()

Get config for writing

property n_non_zero: int

Number of non-zero elements, not counting twice for symmetry!

Returns:

int: number of non-zero elements, not counting twice for symmetry!

class tfConstrainedGauss.solve_id.model_id.ModelID(*args, **kwargs)

Bases: tensorflow.python.keras.engine.training.Model

Model for the Identidy method

__init__(mult_lyr: tfConstrainedGauss.solve_id.model_id.LayerMultPrecCov, **kwargs)

call(input_tensor, training=False)

Wrapper to call the mulitplication layer of precision and covariance matrices

Args:

input_tensor ([type]): Covariance matrix: batch_size x n x n

Returns:

[type]: Product of precision and covariance matrix: vector of length batch_size

classmethod from_config(config)

get_config()

Returns the config of the layer.

A layer config is a Python dictionary (serializable) containing the configuration of a layer. The same layer can be reinstantiated later (without its trained weights) from this configuration.

The config of a layer does not include connectivity information, nor the layer class name. These are handled by Network (one layer of abstraction above).

Note that get_config() does not guarantee to return a fresh copy of dict every time it is called. The callers should make a copy of the returned dict if they want to modify it.

Returns:

Python dictionary.

tfConstrainedGauss.solve_id.solve_id module

class tfConstrainedGauss.solve_id.solve_id.InputsID(n: int, non_zero_idx_pairs: List[Tuple[int, int]], target_cov_mat: <built-in function array>, epochs: int = 100, learning_rate: float = 0.01)

Bases: object

__init__(n: int, non_zero_idx_pairs: List[Tuple[int, int]], target_cov_mat: numpy.array, epochs: int = 100, learning_rate: float = 0.01) → None

convert_mat_non_zero_to_inv_mat(mat_non_zero: numpy.array) → numpy.array

convert_mat_non_zero_to_inv_mat_non_zero(mat_non_zero: numpy.array) → numpy.array

convert_mat_to_mat_non_zero(mat: numpy.array) → numpy.array

epochs: int = 100

learning_rate: float = 0.01

n: int

property n_non_zero

non_zero_idx_pairs: List[Tuple[int, int]]

report()

target_cov_mat: numpy.array

class tfConstrainedGauss.solve_id.solve_id.ResultsID(inputs: tfConstrainedGauss.solve_id.solve_id.InputsID, trained_model: tfConstrainedGauss.solve_id.model_id.ModelID, init_prec_mat_non_zero: <built-in function array>, init_cov_mat_reconstructed: <built-in function array>, learned_prec_mat_non_zero: <built-in function array>, learned_cov_mat: <built-in function array>)

Bases: object

__init__(inputs: tfConstrainedGauss.solve_id.solve_id.InputsID, trained_model: tfConstrainedGauss.solve_id.model_id.ModelID, init_prec_mat_non_zero: numpy.array, init_cov_mat_reconstructed: numpy.array, learned_prec_mat_non_zero: numpy.array, learned_cov_mat: numpy.array) → None

init_cov_mat_reconstructed: numpy.array

init_prec_mat_non_zero: numpy.array

inputs: tfConstrainedGauss.solve_id.solve_id.InputsID

learned_cov_mat: numpy.array

property learned_prec_mat

learned_prec_mat_non_zero: numpy.array

report()

trained_model: tfConstrainedGauss.solve_id.model_id.ModelID

tfConstrainedGauss.solve_id.solve_id.solve_id(inputs: tfConstrainedGauss.solve_id.solve_id.InputsID) → tfConstrainedGauss.solve_id.solve_id.ResultsID

Solve the identity proble

Args:

inputs (InputsID): Inputs

Returns:

ResultsID: Results

Module contents