brainpy.initialize module

This module provides methods to initialize weights. You can access them through brainpy.init.XXX.

Basic Initialization Classes

Initializer

Base Initialization Class.

Regular Initializers

ZeroInit	Zero initializer.
Constant	Constant initializer.
OneInit	One initializer.
Identity	Returns the identity matrix.

Random Initializers

Normal	Initialize weights with normal distribution.
Uniform	Initialize weights with uniform distribution.
TruncatedNormal	Initialize weights with truncated normal distribution.
VarianceScaling
KaimingUniform
KaimingNormal
XavierUniform
XavierNormal
LecunUniform
LecunNormal
Orthogonal	Construct an initializer for uniformly distributed orthogonal matrices.
DeltaOrthogonal	Construct an initializer for delta orthogonal kernels; see arXiv:1806.05393.

Decay Initializers

GaussianDecay	Builds a Gaussian connectivity pattern within a population of neurons, where the weights decay with gaussian function.
DOGDecay	Builds a Difference-Of-Gaussian (dog) connectivity pattern within a population of neurons.

Helper Functions

calculate_gain(nonlinearity[, param])

Return the recommended gain value for the given nonlinearity function.