Dear TMVA team,
I’m dealing with a set of ~4000 BDT classifiers, trained for the signal/background separation in specific angle-momenta ranges. Each time in the beginning of weights application I’m reading all the xml-files with weights (btw, 12 GB), it takes ~30 minutes.
Is it possible to save the configured Reader as a binary object to the .root file? This would accelerate the process. I didn’t find any info in the class reference or User Guide.
Thanks in advance!