search

theogf / ParticleFlow_Exp

3 stars 1 forks source link

Problem with second-order #8

Closed theogf closed 3 years ago

theogf commented 3 years ago

I think there is quite a mess with the second order methods, the notation is very inconsistent and I am not sure about the results anymore. This is needs to be figured out properly

theogf commented 3 years ago

Also I observed that using these preconditionners was sometimes making convergence worse... We need proper experiments for that!