'Word2VecKeyedVectors' object has no attribute 'get_vocab'

[Earl-CS]

Hi!

I'm getting this error when I tried to augment a text. I used a pre-trained embedding called BioWordVec, their github repo is here:

I loaded their vector using gensim

bwv = KeyedVectors.load_word2vec_format('/Thesis/Word Embeddings/BioWordVec_PubMed_MIMICIII_d200.vec.bin', binary=True, limit=1000000)

and used it in WordsEmbAug like this

aug = naw.WordEmbsAug( model_type='word2vec', model= bwv, action="insert")

and when I run

aug_text = aug.augment(text)

I get the error 'Word2VecKeyedVectors' object has no attribute 'get_vocab'

[makcedward]

The class of "model" is custom defined class but not native gensim model class. You may simply use
aug = naw.WordEmbsAug( model_type='word2vec', model_path= '/Thesis/Word Embeddings/BioWordVec_PubMed_MIMICIII_d200.vec.bin', action="insert")

or you can load embeddings into a custom class first and passing it to augmenter

import nlpaug.model.word_embs as nmw
model = nmw.Word2vec()
model.read('/Thesis/Word Embeddings/BioWordVec_PubMed_MIMICIII_d200.vec.bin' , max_num_vector=1000000)

aug = naw.WordEmbsAug( model_type='word2vec', model= model, action="insert")

[Earl-CS]

Thank you!