for testing and deploying your application
for finding and fixing issues
for empowering human code reviews
# Author: Simon Blanke
# Email: [email protected]
# License: MIT License
import pytest
import numpy as np
from gradient_free_optimizers import TreeStructuredParzenEstimators
from ._base_para_test import _base_para_test_func
def objective_function(para):
score = -para["x1"] * para["x1"]
return score
search_space = {"x1": np.arange(-10, 11, 1)}
warm_start_smbo = (
np.array([[-10, -10], [30, 30], [0, 0]]),
np.array([-1, 0, 1]),
)
bayesian_optimizer_para = [
({"gamma_tpe": 0.001}),
({"gamma_tpe": 0.5}),
({"gamma_tpe": 0.9}),
({"warm_start_smbo": None}),
({"warm_start_smbo": warm_start_smbo}),
({"rand_rest_p": 0}),
({"rand_rest_p": 0.5}),
({"rand_rest_p": 1}),
({"rand_rest_p": 10}),
]
pytest_wrapper = ("opt_para", bayesian_optimizer_para)
@pytest.mark.parametrize(*pytest_wrapper)
def test_hill_climbing_para(opt_para):
_base_para_test_func(opt_para, TreeStructuredParzenEstimators)
SimonBlanke /
Gradient-Free-Optimizers
