diffusion_distiller
diffusion_distiller copied to clipboard
How to compute the loglikelihood of the diffusion model?
Is there any way to compute the approximate loglikelihood of the trained diffusion model?