model1 = KeyedVectors.load_word2vec_format('model1.model.bin')
model2 = KeyedVectors.load_word2vec_format('model2.model.bin')
I want t compared vectors between these two model and am using the smart_procrustes_align_gensim code from histwords.
My question is, does loading a model by loading keyedvectors in word2vec format (as above) contain the correct and sufficient information as a full word2vec model so that they can then be used for comparing across models with smart_procrustes_align?
I have checked that syn0 and syn0norm matrices are present, but want to be sure none of these are modified/lost in any way when saving vectors and loading in word2vec format as above.