Okay well in that case - it seems that action function called in FluidPolynomialRegressor.predict
(EDIT: also the .fit
method - maybe they all have the same implementation?) is the source of the errors I’m getting - maybe you can doublecheck the callback implementation for that method? When you get a chance, of course…it’s the weekend, after all!
it is a FluCoMa-wide implementation. so I’ll check but it should be broken everywhere.
So in the above code the degree
and tikhonov
knobs (degreekhoknobs? Sorry, couldn’t help myself) set their respective values and then call a function that looks something like this:
~polyReg.fit(inDataSet, outDataSet,{
~polyReg.predict(questionDataSet, answerDataSet,{
answerDataSet.dump( /* cool stuff happens here*/ )
})
})
This should be safe, right? Or am I totally out to lunch? The errors I’m getting when this gets evaluated too frequently suggest either 1.) .predict
is being called before .fit
is finished and 2). I’m trying to .dump
the answerDataSet before .predict
has finished.
(Perhaps this is going beyond the scope of this thread? Should I open a new thread/bug report on github/something else?)
deffo for github - because this is working for others and as far as I am aware I’m doing the same thing here we’ve done for all objects and nested responding function like you do here should work… so first thing is to try to public objects (any other dataset ones will do) and if you have similar behaviour this is a github bug for sure!
Will do!