pretrained bin and txt have different embeddings

[steveguang]

hi, @rockt, I did something like below,

e2v = gensim.models.KeyedVectors.load_word2vec_format('emoji2vec.bin', binary=True)
print (e2v['😂'])

e2v = gensim.models.KeyedVectors.load_word2vec_format('emoji2vec.txt')
print (e2v['😂'])

It gives 2 different result. Do you know why?