[Bug]: fit terminated unexpected

[yuhorun]

Describe the bug

在使用flaml微调Lightgbm时，迭代53次后，异常终止。我试了2台电脑都是这样，能告诉我是什么原因么？

relation discussions @ #//https://github.com/microsoft/FLAML/discussions/351

Steps to reproduce

我的代码：

settings = { "time_budget": 60606,

"max_iter":60,

"metric": "ap", 
"estimator_list": ["lgbm"],
"task": "classification",
"log_file_name": "houses_experiment.log",
"seed": 7654321,
# "early_stop": True,
"eval_method": "cv",
"n_splits": 3,

} sample_weight = compute_sample_weight(class_weight='balanced', y=Y) automl.fit(X_train=X, y_train=Y, sample_weight=sample_weight, **settings)

Model Used

lgbm

Expected Behavior

No response

Screenshots and logs

终端输出：

[flaml.automl.logger: 11-19 16:16:21] {2442} INFO - at 7700.8s, estimator lgbm's best error=0.2615, best estimator lgbm's best error=0.2615 [flaml.automl.logger: 11-19 16:16:21] {2258} INFO - iteration 51, current learner lgbm [flaml.automl.logger: 11-19 16:22:17] {2442} INFO - at 8056.2s, estimator lgbm's best error=0.2615, best estimator lgbm's best error=0.2615 [flaml.automl.logger: 11-19 16:22:17] {2258} INFO - iteration 52, current learner lgbm [flaml.automl.logger: 11-19 16:31:08] {2442} INFO - at 8587.5s, estimator lgbm's best error=0.2615, best estimator lgbm's best error=0.2615 [flaml.automl.logger: 11-19 16:31:08] {2258} INFO - iteration 53, current learner lgbm

C:\Users\yuhor\Desktop\lgbm>

Additional Information

No response

[thinkall]

@yuhorun , 可能是内存不够导致的。你可以监控一下训练过程中的内存变化吗？谢谢。