SimonBlanke /
Gradient-Free-Optimizers
| Total Complexity | 2 |
| Total Lines | 35 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
| 1 | # Author: Simon Blanke |
||
| 2 | # Email: [email protected] |
||
| 3 | # License: MIT License |
||
| 4 | |||
| 5 | import pytest |
||
| 6 | import numpy as np |
||
| 7 | |||
| 8 | |||
| 9 | from gradient_free_optimizers import HillClimbingOptimizer |
||
| 10 | |||
| 11 | |||
| 12 | def parabola_function(para): |
||
| 13 | loss = para["x"] * para["x"] + para["y"] * para["y"] |
||
| 14 | return -loss |
||
| 15 | |||
| 16 | |||
| 17 | search_space = { |
||
| 18 | "x": np.arange(-10, 11, 1), |
||
| 19 | "y": np.arange(-10, 11, 1), |
||
| 20 | } |
||
| 21 | |||
| 22 | |||
| 23 | def test_epsilon_0(): |
||
| 24 | epsilon = 1 / np.inf |
||
| 25 | |||
| 26 | opt = HillClimbingOptimizer( |
||
| 27 | search_space, initialize={"vertices": 1}, epsilon=epsilon |
||
| 28 | ) |
||
| 29 | opt.search(parabola_function, n_iter=100) |
||
| 30 | |||
| 31 | search_data = opt.search_data |
||
| 32 | scores = search_data["score"].values |
||
| 33 | |||
| 34 | assert np.all(scores == -200) |
||
| 35 |
