Search papers, labs, and topics across Lattice.
1
0
3
Decomposing text prompts into semantic units and using VQA for fine-grained self-reflection dramatically improves image generation quality, especially for complex compositions.
