Clarification about generate_pseudo_pop() parameter

[m-qin]

I’m wondering, in generate_pseudo_pop(), what is the difference between writing something like params = list(xgb_nrounds = 30) vs. params = list(xgb_nrounds = c(10, 20, 30))? What happens if we enter this nrounds parameter as vector versus entering nrounds as a single number? Thanks!

[Naeemkh]

If you enter xgb_nrounds (or any other parameter) as a vector and set max_attempt to any integer more than 1, the function internally will search among those parameters (up to max_attempt) to find the best combination for the covariate balance.
For reproducible results, you can use the value you want to use and set max_attempt = 1.