TypeError in GaussianProcessRegressor.fit() during CBO search about deephyper HOT 4 CLOSED

Hello @theSubsurfaceGuy , the description and screenshot is great. Could you please copy/paste the code from the screenshot so that I do not need to rewrite it? Thank you.

from deephyper.

Hello @Deathn0t, here is the code

from deephyper.problem import HpProblem
from deephyper.evaluator import Evaluator
from deephyper.search.hps import CBO

def objective(hyperparameters):
    x = hyperparameters["x"]
    return (x-5)**2

problem = HpProblem()
problem.add_hyperparameter((1, 10), "x")

evaluator = Evaluator.create(objective, method="process")

search = CBO(problem=problem, evaluator=evaluator, surrogate_model='GP', random_state=42)

results = search.search(max_evals=50)

from deephyper.

@theSubsurfaceGuy I fixed it here: 3a41ac1

You can check out the latest commit from the develop branch, which should be fixed.

Also, DeepHyper performs MAXIMIZATION so if you want to minimize you should return the negative of your true objective.

The results I obtained after using -(x-5)**2 (converging to x=5)

from deephyper.

@Deathn0t thank you very much, I appreciate it.

from deephyper.