In all of the algorithms you have the option to set training params. Specifically the Key and Value as shown below:

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/d87ec76f-2dbe-4032-bc25-9080338ba75f/Screen_Shot_2021-02-02_at_5.52.28_PM.png

Datagran's algorithms are Spark based. For that reason you must refer to the Spark documentation here to find the Params that bets fit your algorithm.