Gents, is it possible to run h2o AutoML multithreaded, but so that they share the same server together or at least so that each thread starts it's own server to train? I see to have an issues here. Running 2 in parallels, when one model finishes faster, it closes the common servers shared and the other instance can't save the resulting model cause the server is already down... Maybe the solution is to make 1 instance of the server and other threads appeal to it for training and saving the models sharing, is this feasible? Thanks!
Gents, is it possible to run h2o AutoML multithreaded, but so that they share the same server together or at least so that each thread starts it's own server to train? I see to have an issues here. Running 2 in parallels, when one model finishes faster, it closes the common servers shared and the other instance can't save the resulting model cause the server is already down... Maybe the solution is to make 1 instance of the server and other threads appeal to it for training and saving the models sharing, is this feasible? Thanks!