We propose a scheme for reduced-precision representation of floating point data on a continuum between IEEE-754 floating point types.
確定! 回上一頁