Questions on the output in the log about flaml HOT 3 CLOSED

Thanks for sharing the feedback about flaml in action. Great to hear that its result is good in your task. To the questions:

1. iter_per_learner means how many models have been tried for each learner. The reason you see iter_per_learner=13 for record_id=2 is that flaml only logs better configs than the previous iters by default, i.e., log_type='better'.
2. trial_time means the time taken to train and evaluate one config in that trial. total_search_time is the total time spent from the beginning of fit().
3. Because log_type='better' by default, config is equal to best_config. If you use log_type='all' instead, all the trials will be logged. And then config corresponds to the config in that iteration, and best_config is the best config so far.
4. flaml will adjust the n_estimators for lightgbm etc. according to the remaining budget and check the time budget constraint and stop in several places. Most of the time that makes fit() stops before the given budget. Occasionally it may run over the time budget slightly. But the log file always contains the best config info and you can recover it using retrain_from_log().

from flaml.

Thank you! All make sense.

from flaml.

Great. If flaml is useful in your application, it will be nice to have your feedback. Feel free to email [email protected] to ask questions or share your feedback.

from flaml.