Tilburg Memory Based Learner: The TiMBL software package is a fast, tree-based implementation of k-nearest neighbor classification. The package includes the IB1, IB2, TRIBL, TRIBL2, and IGTree algorithms, and offers various weighting metrics. Server functionality offered by separate TimblServer wrapper.
TLA-team: State of the art memory-based learning package. Support for several memory-based algorithms, running as a server, and multi-processor systems. Excells at dealing with discrete features (e.g., words). Used as a back-end in many other software projects for several NLP tasks (from morphological splitting to dependency parsing).