ERROR Weka Predictor (3.7) Execute failed: java.lang.NullPointerException

Hello!

I am trying to run LibSVM (weka 3.7), but the predictor shows ERROR Weka Predictor (3.7) Execute failed: java.lang.NullPointerException.
Can anyone help me with this, please?

Hi @mantunes,
can you please post you KNIME.log file? This should tell us more about the error. You can find it under View -> Open KNIME Log.

best,
Gabriel

Log.txt (211.4 KB)
Hello!

I uploaded the log file.

I have experienced that the Weka nodes really do need a lot of memory and especially SVM models are very hungry for it. Besides just using a stronger machine you can try these things:

  • test the whole thing with just 100 items, close all other workflows and restart KNIME before you do to see if it would work at all
  • if it does work with smaller chunks you can think about looping over smaller numbers of lines or try streaming where KNIME would process only several lines at a time (encapsulate the Weka predictor into a meta node)
  • you could see how large you model file is if this is very large that might give you an idea if this whole thing will work at all
  • check how much RAM you have allocated for KNIME
  • try to use the Parquet format to store the data (there is something in the log about loading from a temporary file. Parquet should be more efficient)
  • try to force the Weka node that does the prediction to do everything on the hard drive (instead of the ‘keep small tables in memory’)

Unfortunately my experience is that Weka und especially SVM really do need a lot of power which cannot be easily substituted.

In general check the performance tips on this page:
https://www.knime.com/blog/optimizing-knime-workflows-for-performance

1 Like

Hi!
I tryed just a small partition of the data and it worked! Thanks for your help!

1 Like

Hi there!

Glad it worked with a smaller partition of data. Anyways this is reported and will be checked :wink:

Br,
Ivan

Hi there!

We couldn’t reproduce your issue. Can you maybe share a workflow where we could experience this issue?

Br,
Ivan