Thanks to visit codestin.com
Credit goes to github.com

Skip to content
Open
Overdue by 9 month(s)
Due by February 2, 2025
Last updated Oct 9, 2024
11% complete

laplace-torch v0.3 aims to bring extensive support to diffusion models. These models are challenging due to their (often images) output dimensionality. Considering that the linearized Laplace requires the (n_outputs, n_params)-shaped Jacobians, this can be very challenging (e.g. images have n_outputs = width * height).

All features we aim for are:

  • Efficient, universal, ultimate Jacobian backend
  • Functional Laplace with subset-of-data support
  • Heteroskedastic Laplace
  • Fast last-layer predictive

List view