1 import os 2 # 用於轉換並加載glove預訓練詞向量 3 from gensim.test.utils import datapath, get_tmpfile 4 from gensim.models import KeyedVectors 5 # 將glove轉換為word2vec 6 from gensim.scripts.glove2word2vec import glove2word2vec 7 path=os.getcwd() 8 glove_file=datapath(os.path.join(path, "glove.840B.300d.txt")) 9 tmp_file=get_tmpfile(os.path.join(path,"word2vec.txt")) 10 glove2word2vec(glove_file, tmp_file)
執行完以上代碼后，就在本地生成word2vector形式的預訓練詞向量。執行以上代碼的前提是你下載了glove.840B.300d.txt
下面是加載轉換后的預訓練詞向量
1 # 需要先將glove轉為vector才能用 2 wvmodel = gensim.models.KeyedVectors.load_word2vec_format('word2vec.txt',binary=False, encoding='utf-8')