Cannot compute predictions using "double" only "float"

I cannot generate predictions when I give the TMVA::Experiment::RBDT’s Compute() method an input of std::vector. Indeed, my code fails to compile in this case.

Working code:

TMVA::Experimental::RBDT<> *fBDT1;
fBDT1 = new TMVA::Experimental::RBDT<>(
    "my_bdt", "/path/to/my/model.root");
double prediction = fBDT1->Compute(GetInput())[0];

Where GetInput is defined as std::vector<float> GetInput()
If I try to use std::vector<double> GetInput() instead my code fails to compile with the following message:

required from here
/cvmfs/ error: no matching function for call to ‘TMVA::Experimental::BranchlessJittedForest<float>::Inference(const value_type*, int, bool, __gnu_cxx::__alloc_traits<std::allocator<double>, double>::value_type*)’
          fBackends[i].Inference(&x[0], 1, true, &y[i]);

/cvmfs/ note: candidate: ‘void TMVA::Experimental::BranchlessJittedForest<T>::Inference(const T*, int, bool, T*) [with T = float]’
 void BranchlessJittedForest<T>::Inference(const T *inputs, const int rows, bool layout, T *predictions)
/cvmfs/ note:   no known conversion for argument 1 from ‘const value_type*’ {aka ‘const double*’} to ‘const float*’

Am I not able to use double’s as inputs, if so can this please be fixed?

I guess @moneta can help you.