2 question

Hi guys,

As per the request of my CTO, I am looking for a way to record the logs produced by the MLP learner as it trains on a data set. We require these logs as part of the publication process as evidence that we in fact ran the simulation. 

I have done some googling and found nothing thus far. My team can add this functionality but it would be a pain.

Any help would be appreciated! Especially if anyone can post a WF where this has been accomplished.


and the second question

We are exploring the possibility of using Knime with Raspberry Pi 3.

Note: Raspberry Pi 3 is distributed with Raspbian which includes Java.

Eclipse can be installed from Raspbian repository.

We need to fetch the Knime core so that we can compile, etc.

How do we fetch the Knime core elements, etc?


Hi there,

I’m not sure that setting up KNIME Analytics Platform on a Raspberry Pi would be very easy since it is based on the ARM architecture. You will likely also find that 1GB of RAM is not enough - the default that we use is 2GB for the Java JVM, and you’ll need some left over for the OS. One option may be to use the KNIME Server as the compute power, and either the WebPortal or some REST backed frontend that runs on the Raspberry Pi.

Detailed logs of the MLP output would (I think) require code changes, but if you just need evidence of the parameters used and the output data then the KNIME.log file should be enough. This is also stored on the KNIME Server when a workflow is run.