1、When I invoke built-in algorithm "linear-learner" to train a model , I can use linear.sethyperparameter{} to tune hyper parameters. But in the process, I did not define the arrangement of any hyper parameters, the algorithm can set the tuning range automatically? Because if I create a tuning hyper parameters job, it will require me to provide the range of some parameters to produce different combination of hyper parameters. Thus, can you help me with a explanation that the difference of setting hyper-parameters in creating training job and creating tuning hyper parameters individually?
1、When I invoke built-in algorithm "linear-learner" to train a model , I can use
linear.sethyperparameter{}to tune hyper parameters. But in the process, I did not define the arrangement of any hyper parameters, the algorithm can set the tuning range automatically? Because if I create a tuning hyper parameters job, it will require me to provide the range of some parameters to produce different combination of hyper parameters. Thus, can you help me with a explanation that the difference of setting hyper-parameters in creating training job and creating tuning hyper parameters individually?