Word and multi-sense embedding for Estonian trained on lemmatized etTenTen: Corpus of the Estonian Web. Word embeddings are trained with word2vec. Sense embeddings are trained with SenseGram. Sense inventory is induced ...