Show more
Anyone using things like H2O and similar products? What do you like and dislike?
For non-tree ML models, use Spark mllib. For tree model, use XGBoost. Stay away from H2O.
Thanks for the info. Any specific reasons you dislike the H2o platform? Isn't it just a user friendly interface to spark?
No, it's scalability and reliability are no match for Spark.
keras
Dislike H2o. Spark is nice on distributed data and for pipeline work, but the modeling in mllib kind of stinks. SAS while hated by open sources is really excellent for the modeling, but stay away from it for "deployment"
2024 Presidential Election
12h
482
Weak old Biden
Software Engineering Career
2h
238
why are companies doing “return to office”?
Personal Finance
13h
833
Engagement ring conundrum
India
18h
708
Modi's hate speeches.
For non-tree ML models, use Spark mllib. For tree model, use XGBoost. Stay away from H2O.
Thanks for the info. Any specific reasons you dislike the H2o platform? Isn't it just a user friendly interface to spark?
No, it's scalability and reliability are no match for Spark.