Evaluation results are not added to callbackEnv when metric parameter is left as default

[ruizhuanguw]

Description

I need to specify "metric" or "objective" to get the evaluation results in callBackEnv.evaluation_result_list. If both are left as the default, callBackEnv.evaluation_result_list=[]. This is not consistent with the documentation. https://lightgbm.readthedocs.io/en/latest/Parameters.html#metric

Reproducible example

train_data = lightgbm.Dataset('train_noId.tsv')
valid_data = lightgbm.Dataset('valid_noId.tsv', reference=train_data)

lgbm_params = {
        "boosting_type": "gbdt",
        "label": "name:label",
        "header": "true",
        'num_trees': 10,
    }

separation = lightgbm.train(
    lgbm_params,
    train_set=train_data,
    valid_sets=[valid_data,  train_data],
    valid_names=['Validation', 'Training']
)

Environment info

LightGBM version or commit hash: The behavior is seen in 2.3.0 and 3.2.1.

Command(s) you used to install LightGBM

pip install lightgbm

Additional Comments

[StrikerRUS]

@ruizhuanguw Thanks for using LightGBM and for reporting this issue!

Unfortunately, the piece of code you've provided is not a reproducible example for the described issue. Next time, please try to include all necessary imports, attach all required data and so on to provide a MCVE. This time I've made it for you.

Reproducible example

import lightgbm as lgb
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split

X, y = load_boston(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

train_data = lgb.Dataset(X_train, y_train)
valid_data = lgb.Dataset(X_test, y_test, reference=train_data)

lgbm_params = {
        "boosting_type": "gbdt",
        "objective": "regression",  # commenting out this line makes evals_result empty after training
        'num_trees': 10,
    }

evals_result = {}

separation = lgb.train(
    lgbm_params,
    train_set=train_data,
    valid_sets=[valid_data,  train_data],
    valid_names=['Validation', 'Training'],
    verbose_eval=False,
    callbacks=[lgb.record_evaluation(evals_result)]
)

print(evals_result)

[github-actions[bot]]

This issue has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.