The RapidMiner community is on read-only mode until further notice. Technical support via cases will continue to work as is. For any urgent licensing related requests from Students/Faculty members, please use the Altair academic forum here.

Breakpoint

pantangi_shekarpantangi_shekar Member Posts: 9 Contributor I
edited December 2018 in Help

Can I use a breakpoint to run the model from that particular point??

 

My situation is:

 

Suppose, I have a huge model with n number of joins, m number of filtering and cleaning operators and at the end I have a Neural net.  If I make a small change in one of the join or cleaning operator somewhere in the middle of the model, I had to run the entire model to see the results which is taking 1- 3 mins. Is there any way that I can create a checkpoint in the model so that, I should be able to run the model from that particular point next time instead of running it from the beginning. Is it posisble in Rapidminer????

 

Is that the purpose of Breakpoint? But I dont see it that way. I can observe that at breakpoint the flow stops and shows us the result. 

 

If there is any way to solve this. Please let me know.

 

Thanks in advance.

 

Best,

Raj

Answers

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    Breakpoints are a way to pause your process and review results coming in or going out of an operator. It is especially useful for debugging and such.

     

    You can't select an operator and run your process from there like say a checkpoint, doing that would eat up a lot of memory and RapidMiner is optimized for memory usage. 

  • FBTFBT Member Posts: 106 Unicorn

    As Thomas said, you can't do that with a breakpoint. However, you could use the "Store" operator to store your pre-processing results and then just drag those results into your process. 

  • SGolbertSGolbert RapidMiner Certified Analyst, Member Posts: 344 Unicorn

    I think there is no alternative to the model training time, even if the data is only slightly different, the final model could be VERY different.

     

    That said, it looks as you are spending much of the time in the data prep step. You can easily save some time there by saving an intermediate result with the store operator, especially if you know that the data prep process is constant until a certain point.

     

    Another idea is to try different things in one go, so that you can leave the process running while you drink some coffee :D

Sign In or Register to comment.