How to compute the loglikelihood of the diffusion model?

Hramchenko / diffusion_distiller

🚀 PyTorch Implementation of "Progressive Distillation for Fast Sampling of Diffusion Models(v-diffusion)"

MIT License

217 stars 28 forks source link

Open nuomizai opened 1 year ago

nuomizai commented 1 year ago

Is there any way to compute the approximate loglikelihood of the trained diffusion model?