Rgtemze / HyperDiffusion

Official code repository of "HyperDiffusion: Generating Implicit Neural Fields with Weight-Space Diffusion" @ ICCV 2023
Other
179 stars 16 forks source link

Question for the loss #2

Open SYSUykLin opened 1 year ago

SYSUykLin commented 1 year ago

In the Limitation, ". Our initial experiments with a naive secondary reconstruction loss did not improve performance, but we believe a more sophisticated formulation to encourage the diffusion process to be aware of the surface being represented would improve its generative modeling capabilities". What is the meaning of the naive secondary reconstruction loss?