-
With efficient `fvp` and `ggnvp` #63 it might be nice to add the Hessian-free optimization algorithm of https://www.cs.toronto.edu/~jmartens/docs/Deep_HessianFree.pdf
This algorithm runs a short co…
-
Current layout optimization methods available (Scipy, pyOptSparse's main options) are gradient-based. Paul and I have been working on a gradient free random search approach (#697).
-
We can employ [hyperopt](https://github.com/hyperopt/hyperopt) or [nevergrad](https://github.com/facebookresearch/nevergrad) for optimized hyperparameter search. Grid Search should be considered as op…
-
-
Hi!
I've implemented Conjugate Gradient for Lasagne, and @f0k asked me to put it online so that we can discuss how to integrate it to Lasagne:
https://gist.github.com/neosatrapahereje/b7427ab63dfc05…
-
**Is your feature request related to a problem? Please describe.**
Currently, the README file is not clear about gradient-free optimization and why back-propagation isn't necessary in a gradient-free…
-
**Is your feature request related to a problem? Please describe.**
I'd like to get the gradients for every step during a geometry optimization
**Describe the solution you'd like**
`xtb input.xyz …
-
Prior to #77, the only checks for positive synaptic weights were in `_extract_drive_specs_from_hnn_params()` and have since been removed. We should add a function to drives.py that can be referenced b…
-
### Summary
Expand the function of parameter optimization in DMFF for dynamic properties, such as diffusion coefficient, viscosity, etc. This feature requests,
1. An efficient algorithm for eval…
-
### Description
`find_MAP` is a bit out of style, but it's still a useful bit of functionality in PyMC that could use some love. In particular, I'm thinking about the following upgrades:
1. Allow …