1 paper
Jintao Chen, Aiming Hao, Xiaoqing Chen +6
Pre-trained flow-based models excel at synthesizing complex scenes yet lack a direct mechanism for disentangling and customizing their underlying concepts from one-shot real-world…