SoundCTM: Unifying Score-based and Consistency Models for Full-band Text-to-Sound Generation | Read Paper on Bytez