SOLT (Stochastic Optimization Learning Tool) is a learning tool about gradient descent and stochastic optimization, and in particular the simultaneous perturbation stochastic approximation (SPSA) algorithm.
The SOLT website contains information about the gradient descent and SPSA algorithms together with pseudo-code of the algorithms: https://nanned.github.io/SOLT/.
SOLT contains several Jupyter Notebooks with implementations of the algorithms to experiment with the algorithms by yourself. You can directly use the notebooks in a Binder environment by clicking on the Binder button below:
You can also download the Jupyter Notebooks from the repository and run these in your own environment.
The GNU Affero General Public License v3 (AGPL-3.0) license is used. For more information, please see the included LICENSE.md file.
If you would like to contribute to SOLT in any way, please feel free to create an issue to discuss what you would like to add or change.