Search papers, labs, and topics across Lattice.
1
0
2
2
Forget GAN inversions – now you can steer diffusion models with a dynamically weighted soup of differentiable rewards, including a VQA-based reward for language-vision reasoning, and get SOTA image edits.
