Is it possible log a variable in the objective function?

[yywangvr]

import optuna

def objective(trial):
    x = trial.suggest_float("x", 0, 5)
    y = trial.suggest_float("y", 0, 3)

    v0 = 4 * x ** 2 + 4 * y ** 2
    v1 = (x - 5) ** 2 + (y - 5) ** 2
    v=v0+v1
    return v

Is it possible to log the value of v0 and v1 in some functions such asstudy.trials_dataframe. As I also wish to analyse the intermediate values in the objective functions.

[nzw0301]

Hi @yywangvr, Yes! I think we can store such values in Trial.user_attrs. To do so, we call Trial.set_user_attr during each objective evaluation. For your example

import optuna

def objective(trial):
    x = trial.suggest_float("x", 0, 5)
    y = trial.suggest_float("y", 0, 3)

    v0 = 4 * x ** 2 + 4 * y ** 2
    v1 = (x - 5) ** 2 + (y - 5) ** 2
    trial.set_user_attr("v0", v0)
    trial.set_user_attr("v1", v1)
    v= v0 + v1
    return v

[yywangvr]

Hi @nzw0301 , it works. I really appreciate your sustaining support and help during my using of Optuna!

[nzw0301]

It's my pleasure! I'm glad to hear that I could support your optuna life.