mnrfit takes too long

3 views (last 30 days)
Runcong Kuang
Runcong Kuang on 7 Nov 2022
Commented: Runcong Kuang on 13 Jul 2023
It takes more than two weeks for a single run of mnrfit.
It just reaches 43/100 iter/iterLim in 'nominalFit' after 1.5 weeks.
I am running it on cpu.
When I pause, it's in the function 'nominalFit'.
n=240000 in my case.
To compare, fitcnet causes less than 1 second. Self-designed MLP with 'trainNetwork' costs 450 seconds.
I am pausing the mnrfit now and want to get the results from the first 43 iters.
Is it supposed to end earlier in some condition?

Accepted Answer

Sai Pallav
Sai Pallav on 13 Jul 2023
Please refer to this stack exchange question which tackles a similar situation regarding speed. Furthur, I am adding a documentation of mnrfit, which might help you.
  1 Comment
Runcong Kuang
Runcong Kuang on 13 Jul 2023
Thank you for pointing me to this. I will try the hierarchical mode next time. Whenever you add a documentation on matlab, please let me know. Thanks.

Sign in to comment.

More Answers (0)

Categories

Find more on Startup and Shutdown in Help Center and File Exchange

Products


Release

R2021a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!