Search papers, labs, and topics across Lattice.
1
0
Achieving competitive video super-resolution quality with just 11.25% of the usual trainable parameters, LiteVSR redefines efficiency in adapting frozen diffusion models.
