This package implements some optimization methods used in numerical optimiation tasks for neural networks that aren't usually implemented in deep learning frameworks because of their computational constraints.
Bot releases are hidden (Show)
Added SGD with line search and Conjugate Gradient.
Skipped v0.2.1 since there were some issues with the metadata of the package which had to be fixed.
Published by eugenioLR 4 months ago
The library is now called torch_numopt so that new algorithms can be implemented.
The existing algorithms have also been optimized.