Search papers, labs, and topics across Lattice.
Chinese Academy of Sciences, University of Chinese Academy of Sciences
2
0
5
DiT-Pruning achieves unprecedented image quality retention in Diffusion Transformers, maintaining a CLIP score loss of only 0.001 at 50% sparsity.
LLMs can still be easily fooled by simple prompt rewrites because current adversarial training doesn't adequately cover the data distribution, but a new method using diffusion models closes this gap.
