You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
I have an Euclidean 3-D grid (36x33x5) that I'm trying to fit noiseless. I'm setting the likelihood to the lowest fixed noise possible.
I create a grid using the utils tool to try and reap the benefits of the grid structure:
I then generate the array from it:
This is where my problem begins. My train_x array can also be generated by a simple conversion from numy:
Comparing both with allclose() returns success, so the data is the same.
The issue description is:
My question, is this difference in behaviour somewhat expected? Or is my grid definitions (different densities) hindering the gridGP kernel?
Besides, what additionally is in the "create_data_from_grid" array that causes this difference in behaviour?
For reference, here's the extract from the gridGP class:
Obs.: using an isotropic kernel doesn't make a difference.
The code should be here: https://github.com/ebiga/gpr_aniso_trials/releases/tag/last_pytorch_grid_kernel
Beta Was this translation helpful? Give feedback.
All reactions