The PDB-trained dependency parsing models for Polish is a parsing tool trained on the Polish Dependency Bank (which consists of 22,208 trees and 351,406 tokens). It does part-of-speech tagging, lemmatisation, dependency parsing, and semantic role labeling.