planar_normalizing_flows¶

tfsnippet.layers.planar_normalizing_flows(n_layers=1, w_initializer=<tensorflow.python.ops.init_ops.RandomNormal object>, w_regularizer=None, b_initializer=<tensorflow.python.ops.init_ops.Zeros object>, b_regularizer=None, u_initializer=<tensorflow.python.ops.init_ops.RandomNormal object>, u_regularizer=None, trainable=True, name=None, scope=None)¶

Construct a sequential of :class`PlanarNormalizingFlow`.

Parameters:

n_layers (int) – The number of :class`PlanarNormalizingFlow`.
w_initializer – The initializer for parameter w.
w_regularizer – The regularizer for parameter w.
b_regularizer – The regularizer for parameter b.
b_initializer – The initializer for parameter b.
u_regularizer – The regularizer for parameter u.
u_initializer – The initializer for parameter u.
trainable (bool) – Whether or not the parameters are trainable? (default True)
name (str) – Default name of the variable scope. Will be uniquified. If not specified, generate one according to the class name.
scope (str) – The name of the variable scope.

Returns:

A SequentialFlow: if n_layers > 1, or a PlanarNormalizingFlow if n_layers == 1.

Return type:

SequentialFlow or PlanarNormalizingFlow