gyhandy / Neural-Sim-NeRF

[ECCV 2022] Neural-Sim: Learning to Generate Training Data with NeRF
MIT License
158 stars 8 forks source link

missing conjugate gradient implementation for inverse hessian product? #5

Open hyperzy opened 2 years ago

hyperzy commented 2 years ago

Hey,

It's really a wonderful work on training data generation. I noticed that there is a function get_inverse_hvp_cg & get_inverse_hvp_lissa without definition. Is is intentionally removed?

gyhandy commented 2 years ago

Hi, thanks for your question. We will try to add them in the next version. While they are different approximations, our current implementation is a simple yet efficient approximation version.