Closed andreamola closed 3 years ago
Ok, so to summarize things here, we found one mistake on my part and one tiny bug. First, I was mistakenly under the impression that power was still an attribute of the RBF class, but since 2.0 version it is not the case. I got misled by its presence in some parameter files in the test section. Second, there indeed is a small bug consisting in a commented line (rbf.py:395) that must be uncommented to properly pass additional parameters to the RBFFactory functions. I will go ahead with a PR to fix that line and purge some occurrence of the power parameter in the parameter files around the library.
fix in #226 and #228
There seems to be a problem with the second parameter (power) in most of the RBF kernel functions. Apparently, it is ignored. I have a python scpt and a parameter file to reproduce the issue, if needed.