toooo Looooong optimization/learning.... 8 days and running

fritmore

Been running optimization for 8 days now.... how can I get estimated time to finish the job?

unfortunately I didnt define the log file, but i dont know if it can provide such info.

How do I reuse/save a learned model?

thx

cherokee

Hi fritmore,

depends whether you are runing in gui or from command line. What are you optimizing? Can you provide you proces setup. This make judging easier!

Best regards,
chero

radone

One way is to place an breakpoint somewhere in the looping part of code. Than you can watch and save several types of result.
-radone

fritmore

cherokee wrote:
Hi fritmore, depends whether you are runing in gui or from command line. What are you optimizing? Can you provide you proces setup. This make judging easier! Best regards, chero

Hi chero

am optimizing SVM evolutionary model on labeled continuous values:    (5x30000)data>Optimize>validation(relative split .7,shuffled)>SVM(evolutionary)     

ty
fritmore

fritmore

radone wrote:
One way is to place an breakpoint somewhere in the looping part of code. Than you can watch and save several types of result. -radone

yo radone how should I do that?
How to save current model (or best model so far at such a breakpoint?)
The project is described in the previous post (in the reply to chero)

ty
fritmore

cherokee

Hi fritmore,

i've still some questions:

• Are you working with the gui or without?
• What kind of optimization are you using (Grid, Evolutionary)?
• Which parameters are you optimizing? In which range / which values

Best regards,
chero

P.S.: Your process setup would reveal most of this info   Of course not whether you use the gui.