numpy读取词向量

前言

使用numpy读取词向量。

下载词向量

• GloVe: Global Vectors for Word Representation

读取代码

import numpy as np  

def read(filename):  
  word_to_vec = {}  
  with open(filename) as f:  
      for line in f:  
          first_space_index = line.index(' ')  
          word = line[:first_space_index]  
          values = line[first_space_index + 1:]  
          vector = np.fromstring(values, sep=' ', dtype=np.float16)  
          word_to_vec[word] = vector  
  return word_to_vec

word2vec = read("glove.6B.50d.txt")