Search papers, labs, and topics across Lattice.
1
0
3
Fine-tuning efficient few-step diffusion models no longer requires sacrificing their speed, thanks to a self-distillation approach that preserves inference capabilities.
