Spaces:
Configuration error
Configuration error
File size: 1,999 Bytes
31c7d49 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 | const scalarFloat = new Float32Array(1)
const scalarBits = new Uint32Array(scalarFloat.buffer)
/** Convert one JavaScript number to an IEEE-754 binary16 bit pattern. */
export function numberToFloat16Bits(value: number): number {
scalarFloat[0] = value
const bits = scalarBits[0]
const sign = (bits >>> 16) & 0x8000
let exponent = ((bits >>> 23) & 0xff) - 127 + 15
let mantissa = bits & 0x7fffff
if (exponent <= 0) {
if (exponent < -10) return sign
mantissa = (mantissa | 0x800000) >>> (1 - exponent)
// Round to nearest, ties to even.
const rounded = (mantissa + 0xfff + ((mantissa >>> 13) & 1)) >>> 13
return sign | rounded
}
if (exponent >= 31) {
if (((bits >>> 23) & 0xff) === 0xff && mantissa !== 0) {
return sign | 0x7c00 | Math.max(1, mantissa >>> 13)
}
return sign | 0x7c00
}
mantissa += 0xfff + ((mantissa >>> 13) & 1)
if ((mantissa & 0x800000) !== 0) {
mantissa = 0
exponent += 1
if (exponent >= 31) return sign | 0x7c00
}
return sign | (exponent << 10) | (mantissa >>> 13)
}
export function float32ToFloat16(source: Float32Array): Uint16Array {
const result = new Uint16Array(source.length)
for (let index = 0; index < source.length; index += 1) {
result[index] = numberToFloat16Bits(source[index])
}
return result
}
export function float16BitsToNumber(bits: number): number {
const sign = (bits & 0x8000) !== 0 ? -1 : 1
const exponent = (bits >>> 10) & 0x1f
const mantissa = bits & 0x3ff
if (exponent === 0) {
return mantissa === 0 ? sign * 0 : sign * 2 ** -14 * (mantissa / 1024)
}
if (exponent === 0x1f) return mantissa === 0 ? sign * Infinity : Number.NaN
return sign * 2 ** (exponent - 15) * (1 + mantissa / 1024)
}
export function float16ToFloat32(source: Uint16Array): Float32Array {
const result = new Float32Array(source.length)
for (let index = 0; index < source.length; index += 1) {
result[index] = float16BitsToNumber(source[index])
}
return result
}
|