Vikos is a library for supervised training of parameterized, regression, and classification models.
Design Goals
- Model representations, cost functions, and optimization algorithms can be changed independently of each other.
- Generics: Not committed to a particular data structure for inputs, targets, etc.
- If the design goals above can only be achieved by sacrificing performance, so be it.
Current State
Just starting to get the traits right, by continuously trying new use cases and implementing the learning algorithms. If you are looking for more mature rust libraries in the domain of ML, you might want to check out:
Documentation
Thanks to the folks of docs.rs for building and hosting the documentation!
Contributing
Want to help out? Just create an issue, pull request or contact markus.klein@blue-yonder.com.