search

kaz-Anova / StackNet

StackNet is a computational, scalable and analytical Meta modelling framework
MIT License
1.32k stars 344 forks source link

How to deal with the 'model' folder and 'model2' file everytime we run #44

Closed qq609299174 closed 6 years ago

qq609299174 commented 6 years ago

I follow the steps for Zillow sample. Every time when I tuned the parameters it seems file in model folder changed. Do I have to delete the file or just keep them. Does this affect the results?

goldentom42 commented 6 years ago

Hi, model folder keeps the model fitted with the training data for each fold during each run. Saved models have different ids at each run so keeping them does not affect results. However you may want to delete these files to save disk space. Please keep in mind that if you want to use the predict statement command line( i.e. Java -jar stacknet.jar predict ...) you need to have the corresponding fitted models in the model folder. Hope this helps.

Update: I forgot about model2. This is a recap of all the fitted models and is updated at each train run.

qq609299174 commented 6 years ago

Thanks for your help! Not going to delete them. I have a large disk size XD But another problem here... I am trying to use SklearnknnRegressor. And.. why it is so slow. I stuck at the first round cv for about 15 minutes. I use 16 threads and my CPU utilization was only 30%.

goldentom42 commented 6 years ago

I don't think I'm able to assist you on this particular issue. If you're on windows this may be related to issue #43 It looks like stacknet sometimes hangs on windows though I never encountered the issue but I'm on win8.1 ...

qq609299174 commented 6 years ago

Okay! Thank you solved the problem!