neural_filter_2R.py 2.75 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
"""
NeuralFilter2R
**************

This module implements a trainable all-pole second order filter with real poles using pyTorch


Copyright (c) 2018 Idiap Research Institute, http://www.idiap.ch/

Written by Francois Marelli <Francois.Marelli@idiap.ch>

This file is part of neural_filters.

neural_filters is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License version 3 as
published by the Free Software Foundation.

neural_filters 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 neural_filters. If not, see <http://www.gnu.org/licenses/>.

"""

Francois Marelli's avatar
Francois Marelli committed
28
from . import NeuralFilter
29 30

import torch
31
import numpy as np
32

33 34

class NeuralFilter2R(torch.nn.Module):
35 36 37 38 39 40 41 42 43 44 45
    """
        A trainable second-order all-(real)pole filter :math:`\\frac{1}{1 - P_{1} z^{-1}} \\frac{1}{1 - P_{2} z^{-1}}`

        * **hidden_size** (int) - the size of data vector
        """

    def __init__(self, hidden_size):
        super(NeuralFilter2R, self).__init__()

        self.hidden_size = hidden_size

Francois Marelli's avatar
Francois Marelli committed
46 47
        self.first_cell = NeuralFilter(self.hidden_size)
        self.second_cell = NeuralFilter(self.hidden_size)
48

49
        self.reset_parameters()
Francois Marelli's avatar
Francois Marelli committed
50 51

    def reset_parameters(self, init=None):
52
        if init is None:
Francois Marelli's avatar
Francois Marelli committed
53 54 55
            self.first_cell.reset_parameters(min_modulus=0, max_modulus=0.5)
            self.second_cell.reset_parameters(min_modulus=0.5, max_modulus=1)

56
        elif isinstance(init, tuple):
Francois Marelli's avatar
Francois Marelli committed
57 58 59 60 61 62
            self.first_cell.reset_parameters(init[0])
            self.second_cell.reset_parameters(init[1])
        else:
            self.first_cell.reset_parameters(init)
            self.second_cell.reset_parameters(init)

63
    def __repr__(self):
64
        s = '{name}({hidden_size})'
65 66
        return s.format(name=self.__class__.__name__, **self.__dict__)

M. François's avatar
M. François committed
67
    def forward(self, input_var, hx=(None, None)):
M. François's avatar
M. François committed
68 69
        interm, interm_hidden, first_modulus = self.first_cell(input_var, hx[0])
        output, hidden, second_modulus = self.second_cell(interm, hx[1])
70

M. François's avatar
M. François committed
71
        return output, (interm_hidden, hidden), (first_modulus, second_modulus)
72 73 74 75 76

    @property
    def denominator(self):
        first = self.first_cell.denominator
        second = self.second_cell.denominator
77
        denom = np.zeros((first.shape[0], 3))
78 79 80 81 82 83 84 85
        for i in range(self.hidden_size):
            denom[i] = np.polymul(first[i], second[i])
        return denom

    @property
    def gradients(self):
        first = self.first_cell.gradients
        second = self.second_cell.gradients
86
        return np.concatenate((first, second), axis=1)