H2O AI with own python machine learning model integration with snowflake

Question

i have a few questions regarding H2O AI. As per my understanding, h2o AI powers Auto ML functionality. but need to integrate my own python jupyetr ML model. so my questions are,

1. Can we use H2O AI without Auto ML and with our own python jupyter ML algorithm?
2. If yes, can we integrate that own manual scripted ML with Snowflake?
3. If we can integrate our own scripted ml algorithm with snowflake, what are the advantages of doing it that way? instead of an own manually-created python ML algorithm?

Answer 1

H2O.ai offers a bunch of ML solutions: h2o-3, driverless ai, hydrogen torch to name the main ones.

Driverless AI is AutoML driven, the user has, however, an option to provide a custom recipe (in Python) to customize it. Driverless AI has Snowflake integration.

H2O-3 is a framework that implements a collection of popular ML algorithms. H2O-3 also integrates an AutoML solution utilizing the built-in algos. There is no option to integrate a 3rd party solution into H2O-3 AutoML and to extend H2O-3 algos other than by coding in Java (small Python customizations can be made by providing eg. custom loss function in GBM).