从空格分隔的txt创建numpy数组

例如需要读取的word_vector.txt，每行大致形如下：

以一个单词开头，之后是100维的word2vec数组。

tiger	0.0420839 -0.046227 0.0975495 0.153758 0.156793 -0.0795489 -0.128424 0.0218096 -0.119114 -0.106891 -0.144035 0.0647735 -0.0109524 -0.080547 0.0357672 -0.00600518 -0.0154541 0.0730642 0.189601 0.0752843 -0.293158 0.245404 -0.186248 -0.238262 -0.190642 -0.065447 -0.0754818 -0.0313473 -0.00601225 -0.0144043 -0.0166763 0.0344018 0.0265409 0.0351162 0.172383 -0.118676 -0.0633459 0.0534258 0.0416032 -0.0191747 0.0763039 -0.135899 0.119606 -0.0966475 -0.146213 -0.052332 0.0135779 -0.0465591 0.134781 0.0695233 0.00506461 -0.0387316 0.0129161 0.0958957 -0.0326569 0.17477 0.0438357 -0.10254 -0.0155797 0.0619649 -0.0704648 -0.0985194 0.0365197 0.0338386 0.0167974 -0.0582149 0.220361 -0.0211108 -0.090331 0.216107 -0.11635 -0.0732924 -0.110014 0.0396456 -0.0192266 0.116349 0.0303521 0.0683885 0.0373906 0.00505091 0.0282419 0.102236 0.0771478 0.0926167 0.0406884 0.0399498 -0.134745 -0.102543 0.07766 -0.0937226 0.0567376 0.105321 -0.0160325 0.0732355 -0.0930529 0.0152368 -0.0117502 -0.120294 0.0991484 -0.0874081

这时只需要使用numpy的genfromtxt，注意用usecols指定使用第1到100列来构成数组，这样即可跳过首列的单词。

import numpy as np

X = np.genfromtxt("word_vector.txt", dtype=float, usecols=np.arange(1,101))