gpflow.utilities¶

gpflow.utilities.deepcopy¶

gpflow.utilities.deepcopy(input_module: ~M, memo: Union[Dict[int, Any], NoneType] = None) → ~M[source]¶

Returns a deepcopy of the input tf.Module. To do that first resets the caches stored inside each tfp.bijectors.Bijector to allow the deepcopy of the tf.Module.

Parameters

input_module (~M) – tf.Module including keras.Model, keras.layers.Layer and gpflow.Module.
memo (Optional[Dict[int, Any]]) – passed through to func:copy.deepcopy (see https://docs.python.org/3/library/copy.html).

Return type

~M

Returns

Returns a deepcopy of an input object.

gpflow.utilities.freeze¶

gpflow.utilities.freeze(input_module: ~M) → ~M[source]¶

Returns a deepcopy of the input tf.Module with constants instead of variables and parameters.

Parameters: input_module (~M) – tf.Module or gpflow.Module.
Return type: ~M
Returns: Returns a frozen deepcopy of an input object.

gpflow.utilities.leaf_components¶

gpflow.utilities.leaf_components(input: tensorflow.Module)[source]¶

Parameters: input (tensorflow.Module) –

gpflow.utilities.multiple_assign¶

gpflow.utilities.multiple_assign(module: tensorflow.Module, parameters: Dict[str, tensorflow.Tensor])[source]¶

Multiple assign takes a dictionary with new values. Dictionary keys are paths to the tf.Variable`s or `gpflow.Parameter of the input module.

Parameters

module (tensorflow.Module) – tf.Module.
parameters (Dict[str, tensorflow.Tensor]) – a dictionary with keys of the form “.module.path.to.variable” and new value tensors.

gpflow.utilities.parameter_dict¶

gpflow.utilities.parameter_dict(module: tensorflow.Module) → Dict[str, Union[gpflow.base.Parameter, tensorflow.Variable]][source]¶

Returns a dictionary of parameters (variables) for the tf.Module component. Dictionary keys are relative paths to the attributes to which parameters (variables) assigned to.

class SubModule(tf.Module):

def __init__(self):
self.parameter = gpflow.Parameter(1.0) self.variable = tf.Variable(1.0)

class Module(tf.Module):

def __init__(self):
self.submodule = SubModule()

m = Module() params = parameter_dict(m) # { # “.submodule.parameter”: <parameter object>, # “.submodule.variable”: <variable object> # }

Parameters: module (tensorflow.Module) –
Return type: Dict[str, Union[Parameter, tensorflow.Variable]]

gpflow.utilities.positive¶

gpflow.utilities.positive(lower: Union[float, NoneType] = None, base: Union[str, NoneType] = None) → tensorflow_probability.bijectors.Bijector[source]¶

Returns a positive bijector (a reversible transformation from real to positive numbers).

Parameters

lower (Optional[float]) – overrides default lower bound (if None, defaults to gpflow.config.default_positive_minimum())
base (Optional[str]) – overrides base positive bijector (if None, defaults to gpflow.config.default_positive_bijector())

Return type

tensorflow_probability.bijectors.Bijector

Returns

a bijector instance

gpflow.utilities.print_summary¶

gpflow.utilities.print_summary(module: tensorflow.Module, fmt: str = None)[source]¶

Prints a summary of the parameters and variables contained in a tf.Module.

Parameters

module (tensorflow.Module) –
fmt (Optional[str]) –

gpflow.utilities.read_values¶

gpflow.utilities.read_values(module: tensorflow.Module) → Dict[str, numpy.ndarray][source]¶

Returns a dictionary of numpy values of the module parameters (variables).

Parameters: module (tensorflow.Module) –
Return type: Dict[str, ndarray]

gpflow.utilities.reset_cache_bijectors¶

gpflow.utilities.reset_cache_bijectors(input_module: tensorflow.Module) → tensorflow.Module[source]¶

Recursively finds tfp.bijectors.Bijector-s inside the components of the tf.Module using traverse_component. Resets the caches stored inside each tfp.bijectors.Bijector.

Parameters: input_module (tensorflow.Module) – tf.Module including keras.Model, keras.layers.Layer and gpflow.Module.
Return type: tensorflow.Module
Returns

gpflow.utilities.select_dict_parameters_with_prior¶

gpflow.utilities.select_dict_parameters_with_prior(model: tensorflow.Module) → Dict[str, gpflow.base.Parameter][source]¶

Collects parameters with prior into a dictionary.

Parameters: model (tensorflow.Module) –
Return type: Dict[str, Parameter]

gpflow.utilities.tabulate_module_summary¶

gpflow.utilities.tabulate_module_summary(module: tensorflow.Module, tablefmt: Union[str, NoneType] = None) → str[source]¶

Parameters

module (tensorflow.Module) –
tablefmt (Optional[str]) –

Return type

str

gpflow.utilities.to_default_int¶

gpflow.utilities.to_default_int(x)[source]¶

gpflow.utilities.training_loop¶

gpflow.utilities.training_loop(closure: Callable[[], tensorflow.Tensor], optimizer: Union[tensorflow.optimizers.Optimizer, NoneType] = None, var_list: List[tensorflow.Variable] = None, maxiter=1000.0, compile=False)[source]¶

Simple generic training loop. At each iteration uses a GradientTape to compute the gradients of a loss function with respect to a set of variables.

Parameters

closure (Callable[[], tensorflow.Tensor]) – Callable that constructs a loss function based on data and model being trained
optimizer (Optional[tensorflow.optimizers.Optimizer]) – tf.optimizers or tf.keras.optimizers that updates variables by applying the corresponding loss gradients. Adam is a default optimizer with default settings.
var_list (Optional[List[tensorflow.Variable]]) – List of model variables to be learnt during training
maxiter – Maximum number of

Returns

gpflow.utilities.triangular¶

gpflow.utilities.triangular() → tensorflow_probability.bijectors.Bijector[source]¶

Returns instance of a triangular bijector.

Return type: tensorflow_probability.bijectors.Bijector

gpflow.utilities.bijectors¶

gpflow.utilities.ops¶

gpflow.utilities.ops

gpflow.utilities.utilities¶

gpflow.utilities.utilities