1 paper
Deyuan Liu, Peng Sun, Yansen Han +3
The push for efficient text to image synthesis has moved the field toward one step sampling, yet existing methods still face a three way tradeoff among fidelity, inference speed, a…