The number of binomial terms grows quickly with distance when the number of bitstrings is large. We use this value as a cutoff for doing column renormalization. At worst this number is equal to the dimensionality of the reduced A-matrix, and thus we do not need to compute all terms.
This PR restricts the binomial computation to <= dim(A)
The number of binomial terms grows quickly with distance when the number of bitstrings is large. We use this value as a cutoff for doing column renormalization. At worst this number is equal to the dimensionality of the reduced A-matrix, and thus we do not need to compute all terms.
This PR restricts the binomial computation to <= dim(A)