# Source code for brainpy.initialize.regular_inits

# -*- coding: utf-8 -*-

from brainpy import math as bm, tools
from .base import InterLayerInitializer

__all__ = [
'ZeroInit',
'Constant',
'OneInit',
'Identity',
]

[docs]class ZeroInit(InterLayerInitializer): """Zero initializer. Initialize the weights with zeros. """ def __call__(self, shape, dtype=None): shape = [tools.size2num(d) for d in shape] return bm.zeros(shape, dtype=dtype) def __repr__(self): return self.__class__.__name__
[docs]class Constant(InterLayerInitializer): """Constant initializer. Initialize the weights with the given values. Parameters ---------- value : float, int, bm.ndarray The value to specify. """
[docs] def __init__(self, value=1.): super(Constant, self).__init__() self.value = value
def __call__(self, shape, dtype=None): shape = [tools.size2num(d) for d in shape] return bm.ones(shape, dtype=dtype) * self.value def __repr__(self): return f'{self.__class__.__name__}(value={self.value})'
[docs]class OneInit(Constant): """One initializer. """ pass
[docs]class Identity(InterLayerInitializer): """Returns the identity matrix. This initializer was proposed in (Le, et al., 2015) [1]_. Parameters ---------- value : float The optional scaling factor. Returns ------- shape: tuple of int The weight shape/size. References ---------- .. [1] Le, Quoc V., Navdeep Jaitly, and Geoffrey E. Hinton. "A simple way to initialize recurrent networks of rectified linear units." arXiv preprint arXiv:1504.00941 (2015). """
[docs] def __init__(self, value=1.): super(Identity, self).__init__() self.value = value
def __call__(self, shape, dtype=None): if isinstance(shape, int): shape = (shape,) elif isinstance(shape, (tuple, list)): if len(shape) > 2: raise ValueError(f'Only support initialize 2D weights for {self.__class__.__name__}.') else: raise ValueError(f'Only support shape of int, or tuple/list of int ' f'in {self.__class__.__name__}, but we got {shape}.') shape = [tools.size2num(d) for d in shape] return bm.eye(*shape, dtype=dtype) * self.value def __repr__(self): return f'{self.__class__.__name__}(value={self.value})'