/
get_coefficients.py
44 lines (36 loc) · 1 KB
/
get_coefficients.py
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
"""
get_coefficients
----------------
Get coefficient from model
Summary
-------
Contains all methods associated with getting coefficients from the model
"""
import numpy as np
import math
from keras import backend as K
from keras.layers import Input, Dense, Lambda, concatenate
def get_mixture_coef(output, num_components=24):
"""
Gets output (layer of NN) and converts into proportion, variance and mean
for misture density.
Parameters
----------
output : array
layer of neural network ordered mixture weights (unscaled), variance
(unscaled) and means
num_components : int
number of mixtures
Returns
----------
out_pi : array
weights of mixtures.
out_sigma : array
variance of mixtures.
out_mu : array
mean of mixtures.
"""
out_mu = output[:,:num_components]
out_sigma = output[:,num_components:2*num_components]
out_pi = output[:,2*num_components:]
return out_pi, out_sigma, out_mu