Thanks for the info. I installed htm.core with PyPi before posting, but on loading hotgym.py example from the github I get some errors on predictor.infer().
—> 30 pdf = predictor.infer( tm.getActiveCells() )
31 for n in (1, 5):
32 if pdf[n]:RuntimeError: CHECK FAILED: “dimensions_ != 0” Classifier:
must calllearnbeforeinfer.
Running the hotgym.py file itself in a cloned repo gives me ‘no module named numpy found’ even though I’ve installed numpy normally on my system; I’ll look into a source build rather than PyPi.
Perhaps I’m just missing something fundamental in the code structure; I’ll read up a bit more on how to get htm.core to work.
It looks like I should be able to follow the nupic API documentation while referencing the htm.core API differences for syntax.
