-
Notifications
You must be signed in to change notification settings - Fork 0
/
cGAN_hybrid_extrapolation_testing_results.py
163 lines (121 loc) · 6.57 KB
/
cGAN_hybrid_extrapolation_testing_results.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
159
160
161
162
163
from BeamSFEM.Beam_PlainStress import Beam as beam_linear
from cGAN_hybrid_extrapol import *
import os
import json
import numpy as np
training_keys = [1.0, 4.0, 7.0, 10.0, 13.0, 16.0, 19.0, 22.0, 25.0, 28.0, 31.0]
validation_keys = [2.0, 5.0, 8.0, 11.0, 14.0, 17.0, 20.0, 23.0, 26.0, 29.0]
testing_keys = [3.0, 6.0, 9.0, 12.0, 15.0, 18.0, 21.0, 24.0, 27.0, 30.0, 32.0, 33.0, 34.0, 35.0, 36.0, 37.0, 38.0, 39.0, 40.0]
training_keys = training_keys + validation_keys + testing_keys
# Define SFEM parameters
base_E = float(2.50000000e+09)
E_stdv = 2.57894737e-01
h = 0.4
L = 5.0
rho = 7800
T = 30.0
t = 0.1
CorrLength = [3.05000000e+00]
base_element_size = 0.1
zeta = 0.05
poisson = 0.3
run_SFEM = False
load_SFEM_noise = True
if run_SFEM:
beam = beam_linear(base_E=base_E, E_stdv=E_stdv, h=h, L=L, rho=rho, T=T, t=t,
base_element_size=base_element_size, zeta=zeta, poisson=poisson, CorrLength=CorrLength)
all_SFEM_data = []
training_noise_dict = {}
for load in training_keys:
Us, CoVU = beam.StaticSSFEM(load=load)
SFEM_data = beam.generate_static_samples_SSFEM(10000, Us=Us)[:, -1]
all_SFEM_data.append(SFEM_data)
training_noise_dict[load] = SFEM_data
all_SFEM_data = np.hstack(all_SFEM_data)
min_val = np.min(all_SFEM_data, axis=0)
max_val = np.max(all_SFEM_data, axis=0)
training_keys_ar = np.array(training_keys)
training_keys_norm = 2 * (training_keys_ar - np.min(training_keys_ar, axis=0)) / (np.max(training_keys_ar, axis=0) - np.min(training_keys_ar, axis=0)) - 1.0
to_save_dict = {}
for i, load in enumerate(training_keys):
training_noise_dict[load] = 2 * (training_noise_dict[load] - min_val) / (max_val - min_val) - 1.0
print(training_noise_dict[load].shape)
to_save_dict[str(training_keys_norm[i])] = training_noise_dict[load].tolist()
cwd = os.getcwd()
dict_path = os.path.join(cwd, "data", "training_noise_dict.json")
with open(dict_path, "w") as fp:
json.dump(to_save_dict, fp)
if load_SFEM_noise:
cwd = os.getcwd()
dict_path = os.path.join(cwd, "data", "training_noise_dict.json")
with open(dict_path, "r") as fp:
to_save_dict = json.load(fp)
training_noise_dict = {}
for key in to_save_dict.keys():
training_noise_dict[float(key)] = np.array(to_save_dict[key])
print(training_noise_dict[1.0])
print(training_noise_dict.keys())
training_keys = [1.0, 4.0, 7.0, 10.0, 13.0, 16.0, 19.0, 22.0, 25.0, 28.0, 31.0]
validation_keys = [2.0, 5.0, 8.0, 11.0, 14.0, 17.0, 20.0, 23.0, 26.0, 29.0]
testing_keys = [3.0, 6.0, 9.0, 12.0, 15.0, 18.0, 21.0, 24.0, 27.0, 30.0, 32.0, 33.0, 34.0, 35.0, 36.0, 37.0, 38.0,
39.0, 40.0]
testing_keys_normal = [3.0, 6.0, 9.0, 12.0, 15.0, 18.0, 21.0, 24.0, 27.0, 30.0]
testing_keys_extrapol = [32.0, 33.0, 34.0, 35.0, 36.0, 37.0, 38.0, 39.0, 40.0]
cwd = os.getcwd()
name = "nonlin_Us.json"
file_path = os.path.join(cwd, "data", name)
training_data_norm, training_codes_norm, validation_data, validation_codes_norm, \
testing_data, testing_codes_norm, train_min, train_max = load_real_samples_min_n_max(file_path, training_keys, validation_keys, testing_keys)
print(validation_data[:, -1])
print(validation_codes_norm)
validation_codes_norm_new = []
for datum in validation_data:
code = datum[-1]
validation_codes_norm_new.append(code)
validation_codes_norm_new = list(set(validation_codes_norm_new))
validation_codes_norm = validation_codes_norm_new
all_keys = np.hstack((training_codes_norm, validation_codes_norm, testing_codes_norm)).tolist()
all_keys = sorted(all_keys)
new_training_noise_dict = {}
for i, key in enumerate(sorted(training_noise_dict.keys())):
new_training_noise_dict[all_keys[i]] = training_noise_dict[key]
training_noise_dict = new_training_noise_dict
val_kernels = calculate_validation_data_kernels(validation_data, validation_codes_norm)
noise_dim = 1
sample_dim = 1
code_dim = 1
cwd = os.getcwd()
name = "nonlin_Us.json"
file_path = os.path.join(cwd, "data", name)
testing_avg_kl_diverg = 0
max_KL = -1.0
path = os.path.join(cwd, "saved_models", "CGAN_hybrid_generator_extrapol_results.h5")
model = load_model(path)
model.summary()
for i, test_code in enumerate(testing_codes_norm):
if testing_keys[i] in testing_keys_normal:
print(testing_keys[i])
kl_diverg, reals, fakes = test_best_generator_distribs(path, sample_dim=sample_dim, code_dim=code_dim, noise_dim=noise_dim, testing_data=testing_data,
testing_codes_norm_value=test_code, validation_codes_norm=testing_codes_norm, train_min=train_min, train_max=train_max, training_noise_dict=training_noise_dict)
testing_avg_kl_diverg += kl_diverg
testing_avg_kl_diverg /= len(testing_keys_normal)
print("Mean Testing KL Divegence: ", testing_avg_kl_diverg)
testing_avg_kl_diverg_extrapol = 0
for i, test_code in enumerate(testing_codes_norm):
if testing_keys[i] in testing_keys_extrapol:
print(testing_keys[i], test_code)
kl_diverg, reals, fakes = test_best_generator_distribs(path, sample_dim=sample_dim, code_dim=code_dim, noise_dim=noise_dim, testing_data=testing_data,
testing_codes_norm_value=test_code, validation_codes_norm=testing_codes_norm, train_min=train_min, train_max=train_max, training_noise_dict=training_noise_dict)
testing_avg_kl_diverg_extrapol += kl_diverg
testing_avg_kl_diverg_extrapol /= len(testing_keys_extrapol)
print("Mean Testing KL Divegence extrapolation: ", testing_avg_kl_diverg_extrapol)
val_avg_kl_diverg = 0
for val_code in validation_codes_norm:
kl_diverg, reals, fakes = test_best_generator_distribs(path, sample_dim=sample_dim, code_dim=code_dim, noise_dim=noise_dim, testing_data=validation_data,
testing_codes_norm_value=val_code, validation_codes_norm=validation_codes_norm, train_min=train_min, train_max=train_max, training_noise_dict=training_noise_dict)
val_avg_kl_diverg += kl_diverg
val_avg_kl_diverg /= len(validation_keys)
print("Mean Validation KL Divegence: ", val_avg_kl_diverg)
for train_code in training_codes_norm:
kl_diverg, reals, fakes = test_best_generator_distribs(path, sample_dim=sample_dim, code_dim=code_dim, noise_dim=noise_dim, testing_data=training_data_norm,
testing_codes_norm_value=train_code, validation_codes_norm=training_codes_norm, train_min=train_min, train_max=train_max, training_noise_dict=training_noise_dict)