3 datasets found

Distributional language models for Polish trained on different corpora (KGR10, NKJP, Wikipedia).

Evaluation tools (WBST, HWBST, EWBST) for word embedding models used to assess and compare the usefulness of different word embeddings

Model skip gram with vectors of length 100. Trained on kgr 10, a corpora with over 4 billion tokens. Data preprocessing involved segmentation, lemmatization and mophosyntactic...