MAX_VALUE

Maximum positive finite value a half-precision float may have.