(Translated by https://www.hiragana.jp/)
[2312.06059] CONFORM: Contrast is All You Need For High-Fidelity Text-to-Image Diffusion Models