ALESS_2020_A_M3/new.py at main · sitahara/ALESS_2020_A_M3 · GitHub

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
from keras.layers import Input,Dense,Conv2D,MaxPooling2D,Flatten,BatchNormalization,Reshape,LeakyReLU,Flatten
from keras.models import Model
from keras.utils import to_categorical

import numpy as numpy
import keras
import numpy as np
import matplotlib.pyplot as plt

import joblib

image_shape=(28,28)
in_shape=(20,28)

def build_generator_1(middle_dim1=256,middle_dim2=128,middle_dim3=32):
	inputs=Input(shape=in_shape, name="Input")
	mid1=Flatten()(inputs)#(20*28)*1
	middle=Dense(middle_dim1,activation="relu",name="1st")(mid1)#
	middle=Dense(middle_dim2,activation="relu",name="2nd")(middle)
	core=Dense(middle_dim3,activation="relu",name="3rd")(middle)
	middle=Dense(middle_dim2,activation="relu",name="4th")(core)
	middle=Dense(224,activation="linear",name="5th")(middle)#このあたりを変える
	outputs=Reshape((8,28))(middle)
	generator=Model(inputs=inputs,outputs=outputs,name="Generator")
	return generator

def build_generator_2(middle_dim1=256,middle_dim2=128,middle_dim3=32):
	inputs=Input(shape=in_shape, name="Input")
	mid1=Flatten()(inputs)#(20*28)*1
	middle=Dense(middle_dim1,activation="relu",name="1st")(mid1)
	middle=Dense(middle_dim2*2,activation="relu",name="2nd")(middle)
	middle=Dense(middle_dim2,activation="relu",name="3rd")(middle)
	core=Dense(middle_dim3,activation="relu",name="4th")(middle)
	middle=Dense(middle_dim2,activation="relu",name="5th")(core)
	middle=Dense(middle_dim2*2,activation="relu",name="6th")(middle)
	middle=Dense(224,activation="linear",name="7th")(middle)#このあたりを変える
	outputs=Reshape((8,28))(middle)
	generator=Model(inputs=inputs,outputs=outputs,name="Generator")
	return generator

def build_generator_3(middle_dim1=256,middle_dim2=128,middle_dim3=32):
	inputs=Input(shape=in_shape, name="Input")
	mid1=Flatten()(inputs)#(20*28)*1
	middle=Dense(middle_dim1,activation="relu",name="1st")(mid1)
	middle=Dense(middle_dim2,activation="relu",name="2nd")(middle)
	middle=Dense(middle_dim2,activation="relu",name="3rd")(middle)
	middle=Dense(middle_dim2,activation="relu",name="4th")(middle)
	core=Dense(middle_dim3,activation="relu",name="5th")(middle)
	middle=Dense(middle_dim2,activation="relu",name="6th")(core)
	middle=Dense(middle_dim2,activation="relu",name="7th")(middle)
	middle=Dense(middle_dim2,activation="relu",name="8th")(middle)
	middle=Dense(224,activation="linear",name="9th")(middle)#このあたりを変える
	outputs=Reshape((8,28))(middle)
	generator=Model(inputs=inputs,outputs=outputs,name="Generator")
	return generator

def build_generator_4(middle_dim1=256,middle_dim2=128,middle_dim3=32):
	inputs=Input(shape=in_shape, name="Input")
	mid1=Flatten()(inputs)#(20*28)*1
	middle=Dense(middle_dim3,activation="relu",name="1st")(mid1)
	middle=Dense(middle_dim2,activation="relu",name="2nd")(middle)
	core=Dense(middle_dim1,activation="relu",name="3rd")(middle)
	middle=Dense(middle_dim1*3,activation="relu",name="4th")(core)
	middle=Dense(224,activation="linear",name="5th")(middle)#このあたりを変える
	outputs=Reshape((8,28))(middle)
	generator=Model(inputs=inputs,outputs=outputs,name="Generator")
	return generator

def build_generator_5(middle_dim1=256,middle_dim2=128,middle_dim3=32):
	inputs=Input(shape=in_shape, name="Input")
	mid1=Flatten()(inputs)#(20*28)*1
	middle=Dense(middle_dim3,activation="relu",name="1st")(mid1)
	middle=Dense(middle_dim2,activation="relu",name="2nd")(middle)
	middle=Dense(middle_dim2*2,activation="relu",name="3rd")(middle)
	core=Dense(middle_dim1,activation="relu",name="4th")(middle)
	middle=Dense(middle_dim1*2,activation="relu",name="5th")(core)
	middle=Dense(middle_dim1*3,activation="relu",name="6th")(middle)
	middle=Dense(224,activation="linear",name="7th")(middle)#このあたりを変える
	outputs=Reshape((8,28))(middle)
	generator=Model(inputs=inputs,outputs=outputs,name="Generator")
	return generator


def build_discriminator():
	inputs=Input(shape=image_shape)

	middle=Flatten()(inputs)
	middle=Dense(1000)(middle)
	middle=LeakyReLU(alpha=0.2)(middle)
	middle=Dense(1000)(middle)
	middle=LeakyReLU(alpha=0.2)(middle)
	middle=Dense(1000)(middle)
	middle=LeakyReLU(alpha=0.2)(middle)
	outputs=Dense(20,activation="sigmoid")(middle)#本物の0~9か、偽物の0~9か

	discriminator=Model(inputs=inputs,outputs=outputs,name="Discriminator")
	return discriminator