forked from zsoltzombori/generative_modeling
-
Notifications
You must be signed in to change notification settings - Fork 0
/
squaresample.py
161 lines (123 loc) · 3.75 KB
/
squaresample.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
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.offsetbox import OffsetImage, AnnotationBbox
from matplotlib.patches import Ellipse
import scipy.stats
import os
import model_IO
import params
import data
import vis
def cdf(x):
return scipy.stats.norm.cdf(x, loc=0, scale=1)
def icdf(x):
return scipy.stats.norm.isf(1 - x, loc=0, scale=1)
n = 100
xs = []
ys = []
"""
for r in np.linspace(0, 2, 10):
for i in range(n+1):
x = np.cos(i*2*np.pi/n) * r
y = np.sin(i*2*np.pi/n) * r
xs.append(cdf(x))
ys.append(cdf(y))
plt.plot(xs, ys)
plt.show()
"""
m = 40
n = m ** 2
for u in np.linspace(0, 1, m+2)[1:-1]:
for v in np.linspace(0, 1, m+2)[1:-1]:
xs.append(icdf(u))
ys.append(icdf(v))
#print(len(xs))
#print(len(ys))
"""
latentpoints = np.zeros((len(xs), 2))
for i in range(len(xs)):
latentpoints[i][0] = xs[i]
latentpoints[i][1] = ys[i]
"""
#latentpoints = np.array(zip(xs, ys))
latentpoints = np.array([xs, ys]).T
#print(latentpoints)
args = params.getArgs()
modelDict = model_IO.load_autoencoder(args)
encoder = modelDict.encoder
generator = modelDict.generator
data_object = data.load(args.dataset, shape=args.shape, color=args.color)
((x_train, y_train), (x_test, y_test)) = data_object.get_data(args.trainSize, args.testSize)
args.original_shape = x_train.shape[1:]
args.original_size = np.prod(args.original_shape)
latims = x_train[:1000]
labels = y_train[:1000]
z_sampled, z_mean, z_logvar = encoder.predict(latims, args.batch_size)
images = generator.predict(latentpoints, args.batch_size)
inum = len(latims)
ells = [Ellipse(xy = z_mean[i],
width = 2 * np.exp(z_logvar[i][0]),
height = 2 * np.exp(z_logvar[i][1]))
for i in range(inum)]
#print(images.shape)
#vis.plotImages(images, m, m, "grid")
"""
for p, img in zip(latentpoints, images):
plt.imshow(img[:,:,0], extent = [p[0]-0.1, p[0]+0.1, p[1]-0.1, p[1]+0.1])
plt.show()
"""
def getImage(i, zoom=0.05):
return OffsetImage(images[i][:,:,0], zoom=zoom)
#fig = plt.figure(10, 10)
fig, ax = plt.subplots(subplot_kw = {'aspect' : 'equal'})
ax.scatter(xs, ys, s = None, c = "white")
for i in range(n):
ab = AnnotationBbox(getImage(i), (latentpoints[i][0], latentpoints[i][1]), frameon=False)
ax.add_artist(ab)
blu = []
g = []
r = []
c = []
m = []
y = []
bla = []
o = []
t = []
br = []
for i in range(inum):
ax.add_artist(ells[i])
ells[i].set_clip_box(ax.bbox)
ells[i].set_alpha(0.5)
if labels[i] == 0:
ells[i].set_facecolor('blue')
blu.append(ells[i])
elif labels[i] == 1:
ells[i].set_facecolor('green')
g.append(ells[i])
elif labels[i] == 2:
ells[i].set_facecolor('red')
r.append(ells[i])
elif labels[i] == 3:
ells[i].set_facecolor('cyan')
c.append(ells[i])
elif labels[i] == 4:
ells[i].set_facecolor('magenta')
m.append(ells[i])
elif labels[i] == 5:
ells[i].set_facecolor('yellow')
y.append(ells[i])
elif labels[i] == 6:
ells[i].set_facecolor('black')
bla.append(ells[i])
elif labels[i] == 7:
ells[i].set_facecolor('orange')
o.append(ells[i])
elif labels[i] == 8:
ells[i].set_facecolor('teal')
t.append(ells[i])
else:
ells[i].set_facecolor('brown')
br.append(ells[i])
ax.legend((blu[0], g[0], r[0], c[0], m[0], y[0], bla[0], o[0], t[0], br[0]),
(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), loc="best")
plt.savefig("gridnumbers.png", dpi = 1000)