Discriminative Probing and Tuning for Text-to-Image Generation | Read Paper on Bytez