Use 'lightgbm_tuner:lgbm_params' in https://gist.github.com/smly/584e4b9eabacecc7c102d47de4066a2f

lightgbm_tuner_with_mlflow.py

import json

import mlflow
import numpy as np
import sklearn.datasets
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split

import optuna
import optuna.integration.lightgbm as lgb
from optuna.integration.lightgbm_tuner.optimize import DEFAULT_LIGHTGBM_PARAMETERS


def mlflow_callback(study, trial):
    trial_value = trial.value if trial.value is not None else float("nan")
    trial_runtime = float("nan")
    if trial.datetime_start and trial.datetime_complete:
        trial_runtime = (trial.datetime_complete - trial.datetime_start).total_seconds()

    with mlflow.start_run(run_name=study.study_name):
        lgbm_params = json.loads(trial.user_attrs["lightgbm_tuner:lgbm_params"])
        tuned_params = {k:lgbm_params[k] for k in DEFAULT_LIGHTGBM_PARAMETERS.keys()}
        mlflow.log_params(tuned_params)
        mlflow.log_params({"step_name": trial.user_attrs["lightgbm_tuner:step_name"]})
        mlflow.log_metrics({
            "trial_number": trial.number,
            "elapsed_time": trial_runtime,
            "mean_squared_error": trial_value,
        })


if __name__ == "__main__":
    data, target = sklearn.datasets.load_breast_cancer(return_X_y=True)
    train_x, val_x, train_y, val_y = train_test_split(data, target, test_size=0.25)
    dtrain = lgb.Dataset(train_x, label=train_y)
    dval = lgb.Dataset(val_x, label=val_y)

    params = {
        "objective": "binary",
        "metric": "binary_logloss",
        "verbosity": -1,
        "boosting_type": "gbdt",
    }
    study = optuna.create_study(study_name="lightgbm_tuner_breast_cancer", direction="minimize")

    model = lgb.train(
        params, dtrain, valid_sets=[dtrain, dval], verbose_eval=100, early_stopping_rounds=100, optuna_callbacks=[mlflow_callback]
    )

    prediction = np.rint(model.predict(val_x, num_iteration=model.best_iteration))
    accuracy = accuracy_score(val_y, prediction)

    best_params = model.params
    print("Best params:", best_params)
    print("  Accuracy = {}".format(accuracy))
    print("  Params: ")
    for key, value in best_params.items():
        print("    {}: {}".format(key, value))