nonlinearities.py 1.73 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
# Copyright(C) 2013-2018 Max-Planck-Society
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik
# and financially supported by the Studienstiftung des deutschen Volkes.

19
from __future__ import absolute_import, division, print_function
20

21
from .compat import *
22
from .sugar import exp, full, tanh
23 24


25
class Linear(object):
26 27 28 29
    def __call__(self, x):
        return x

    def derivative(self, x):
Martin Reinecke's avatar
step 1  
Martin Reinecke committed
30
        return full(x.domain, 1.)
31

32
    def hessian(self, x):
Martin Reinecke's avatar
step 1  
Martin Reinecke committed
33
        return full(x.domain, 0.)
34

35 36 37 38 39 40 41 42

class Exponential(object):
    def __call__(self, x):
        return exp(x)

    def derivative(self, x):
        return exp(x)

43 44 45
    def hessian(self, x):
        return exp(x)

46 47 48 49 50 51 52 53

class Tanh(object):
    def __call__(self, x):
        return tanh(x)

    def derivative(self, x):
        return (1. - tanh(x)**2)

54 55
    def hessian(self, x):
        return - 2. * tanh(x) * (1. - tanh(x)**2)
56

Martin Reinecke's avatar
Martin Reinecke committed
57

58
class PositiveTanh(object):
59 60 61 62 63
    def __call__(self, x):
        return 0.5 * tanh(x) + 0.5

    def derivative(self, x):
        return 0.5 * (1. - tanh(x)**2)
64 65 66

    def hessian(self, x):
        return - tanh(x) * (1. - tanh(x)**2)