|
1
|
|
|
import time |
|
2
|
|
|
import pytest |
|
3
|
|
|
import numpy as np |
|
4
|
|
|
import pandas as pd |
|
5
|
|
|
|
|
6
|
|
|
from hyperactive import Hyperactive |
|
7
|
|
|
|
|
8
|
|
|
|
|
9
|
|
|
def func1(): |
|
10
|
|
|
pass |
|
11
|
|
|
|
|
12
|
|
|
|
|
13
|
|
|
def func2(): |
|
14
|
|
|
pass |
|
15
|
|
|
|
|
16
|
|
|
|
|
17
|
|
|
class class1: |
|
18
|
|
|
def __init__(self): |
|
19
|
|
|
pass |
|
20
|
|
|
|
|
21
|
|
|
|
|
22
|
|
|
class class2: |
|
23
|
|
|
def __init__(self): |
|
24
|
|
|
pass |
|
25
|
|
|
|
|
26
|
|
|
|
|
27
|
|
|
def class_f1(): |
|
28
|
|
|
return class1 |
|
29
|
|
|
|
|
30
|
|
|
|
|
31
|
|
|
def class_f2(): |
|
32
|
|
|
return class2 |
|
33
|
|
|
|
|
34
|
|
|
|
|
35
|
|
|
def numpy_f1(): |
|
36
|
|
|
return np.array([0, 1]) |
|
37
|
|
|
|
|
38
|
|
|
|
|
39
|
|
|
def numpy_f2(): |
|
40
|
|
|
return np.array([1, 0]) |
|
41
|
|
|
|
|
42
|
|
|
|
|
43
|
|
|
search_space = { |
|
44
|
|
|
"x0": list(range(-3, 3)), |
|
45
|
|
|
"x1": list(np.arange(-1, 1, 0.001)), |
|
46
|
|
|
"string0": ["str0", "str1"], |
|
47
|
|
|
"function0": [func1, func2], |
|
48
|
|
|
"class0": [class_f1, class_f2], |
|
49
|
|
|
"numpy0": [numpy_f1, numpy_f2], |
|
50
|
|
|
} |
|
51
|
|
|
|
|
52
|
|
|
|
|
53
|
|
|
def objective_function(opt): |
|
54
|
|
|
score = -opt["x1"] * opt["x1"] |
|
55
|
|
|
return score |
|
56
|
|
|
|
|
57
|
|
|
|
|
58
|
|
|
def test_warm_start_0(): |
|
59
|
|
|
hyper0 = Hyperactive() |
|
60
|
|
|
hyper0.add_search(objective_function, search_space, n_iter=15) |
|
61
|
|
|
hyper0.run() |
|
62
|
|
|
|
|
63
|
|
|
best_para0 = hyper0.best_para(objective_function) |
|
64
|
|
|
|
|
65
|
|
|
hyper1 = Hyperactive() |
|
66
|
|
|
hyper1.add_search( |
|
67
|
|
|
objective_function, |
|
68
|
|
|
search_space, |
|
69
|
|
|
n_iter=15, |
|
70
|
|
|
initialize={"warm_start": [best_para0]}, |
|
71
|
|
|
) |
|
72
|
|
|
hyper1.run() |
|
73
|
|
|
|
|
74
|
|
|
|
|
75
|
|
|
def test_warm_start_1(): |
|
76
|
|
|
hyper0 = Hyperactive(distribution="pathos") |
|
77
|
|
|
hyper0.add_search(objective_function, search_space, n_iter=15, n_jobs=2) |
|
78
|
|
|
hyper0.run() |
|
79
|
|
|
|
|
80
|
|
|
best_para0 = hyper0.best_para(objective_function) |
|
81
|
|
|
|
|
82
|
|
|
hyper1 = Hyperactive() |
|
83
|
|
|
hyper1.add_search( |
|
84
|
|
|
objective_function, |
|
85
|
|
|
search_space, |
|
86
|
|
|
n_iter=15, |
|
87
|
|
|
initialize={"warm_start": [best_para0]}, |
|
88
|
|
|
) |
|
89
|
|
|
hyper1.run() |
|
90
|
|
|
|
|
91
|
|
|
|
|
92
|
|
|
def test_warm_start_2(): |
|
93
|
|
|
hyper0 = Hyperactive() |
|
94
|
|
|
hyper0.add_search(objective_function, search_space, n_iter=15) |
|
95
|
|
|
hyper0.run() |
|
96
|
|
|
|
|
97
|
|
|
best_para0 = hyper0.best_para(objective_function) |
|
98
|
|
|
|
|
99
|
|
|
hyper1 = Hyperactive(distribution="pathos") |
|
100
|
|
|
hyper1.add_search( |
|
101
|
|
|
objective_function, |
|
102
|
|
|
search_space, |
|
103
|
|
|
n_iter=15, |
|
104
|
|
|
n_jobs=2, |
|
105
|
|
|
initialize={"warm_start": [best_para0]}, |
|
106
|
|
|
) |
|
107
|
|
|
hyper1.run() |
|
108
|
|
|
|
|
109
|
|
|
|
|
110
|
|
|
def test_warm_start_3(): |
|
111
|
|
|
hyper0 = Hyperactive(distribution="pathos") |
|
112
|
|
|
hyper0.add_search(objective_function, search_space, n_iter=15, n_jobs=2) |
|
113
|
|
|
hyper0.run() |
|
114
|
|
|
|
|
115
|
|
|
best_para0 = hyper0.best_para(objective_function) |
|
116
|
|
|
|
|
117
|
|
|
hyper1 = Hyperactive(distribution="pathos") |
|
118
|
|
|
hyper1.add_search( |
|
119
|
|
|
objective_function, |
|
120
|
|
|
search_space, |
|
121
|
|
|
n_iter=15, |
|
122
|
|
|
n_jobs=2, |
|
123
|
|
|
initialize={"warm_start": [best_para0]}, |
|
124
|
|
|
) |
|
125
|
|
|
hyper1.run() |
|
126
|
|
|
|
SimonBlanke /
Hyperactive
