Would be much faster than current implementation of processing each sentence pair individually.
Would be much faster than current implementation of processing each sentence pair individually.