English and Portuguese CBOW Models from Europarl Corpus, version 7, using FastText with Subwords Option

The models were trained using FastText, model CBOW, 40 epochs, and subwords. Each *.BIN file has its *.VEC file with the vocabulary ordered by frequency. The *.BIN file can return a vector to represent an out-of-vocabulary (OOV) word if the necessary parts of the OOV word were used in training. FastText and Gensim can use these files. The English and Portuguese models are identified in the file name,  _en_  and _pt_ respectively.

An Excel file has the neighborhood changes of some selected words during training on each epoch. A previous exercise to find words with more than one meaning.