Hi Christoph, I think same topic was discussed in github issue. You can use Use avg=True option to avoid the problem. I hope the URL would be help for you. https://github.com/rdkit/rdkit/issues/1527
Best regards, Taka 2019年4月5日(金) 18:37 Christoph Hillisch <christoph.hilli...@univie.ac.at>: > Hello all, > > I use RDKit to calculate descriptors, which I use to train a random forest > model in scikit-learn. > Since I do not scale my training data, I run into the problem that the > descriptor Ipc may contain huge figures (1E+50), which then are too large > for the data type float32 used in sklearn. > > Is there a way of making sure the value of this descriptor fits in > float32, without scaling my data? > Otherwise I’d probably have to remove this descriptor from my model. > > Best regards, > Christoph > > _______________________________________________ > Rdkit-discuss mailing list > Rdkit-discuss@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/rdkit-discuss >
_______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
