Linfa's 0.6.0 release removes the mandatory dependency on external BLAS libraries (such as
intel-mkl) by using a pure-Rust linear algebra library. It also adds the Naive Multinomial Bayes and Follow The Regularized Leader algorithms. Additionally, the
AsTargets trait has been separated into
Linfa's 0.5.0 release adds initial support for the OPTICS algorithm, multinomials logistic regression, and the family of nearest neighbor algorithms. Furthermore, we have improved documentation and introduced hyperparameter checking to all algorithms.
Linfa's 0.4.0 release introduces four new algorithms, improves documentation of the ICA and K-means implementations, adds more benchmarks to K-Means and updates to ndarray's 0.14 version.
In this release of Linfa the documentation is extended, new examples are added and the functionality of datasets improved. No new algorithms were added.
I'm happy to announce that Linfa finally gets its own website. We are currently trying to improve the documentation in various places and a website should serve as a starting point for anyone interested in Linfa.
Linfa 0.3.0 concentrates on polishing the existing implementation and adds only three new algorithms to the crowd. A new feature system is introduced, which allows the selection of the BLAS/LAPACK backend in the base-crate. The
Dataset interface is polished and follows the
ndarray model more closely. The new
linfa-datasets crate gives easier access to sample datasets and can be used for testing.
Linfa 0.2.1 amends changes to the feature system, which made it impossible to release 0.2.0 on
This release of Linfa introduced 9 new implementations and a couple of changes to the APIs. Travis support for FOSS projects was dropped, so we were forced to switch to Github Actions and we introduced a couple of traits to represent different classes of algorithms in a better way.