如何更改numpy数组的输出

Question

我有一个这样的数据集

我尝试使用以下代码对词汇进行编码

def norm_vector(sentence, model, stopwords):
    vecs = [model[word.lower()] for word in word_tokenize(sentence) if word not in stopwords]  
    norm_vecs = [vec / np.linalg.norm(vec) for vec in vecs if np.linalg.norm(vec) > 0]
    sent_vecs = np.mean(norm_vecs, axis=0)
    return sent_vecs  

vecs_train_LV = [norm_vector(sentence, model_level, sw_indo) for sentence in data.data]
vecs_train_LV = np.array(vecs_train_LV)
vecs_train_LV

这段代码的输出如下：

array([array([-4.35138009e-02, -2.87008341e-02,  9.86183342e-03, -8.87360424e-02,
        1.83405634e-02, -6.50617108e-02, -6.65896460e-02,  8.60413313e-02,
       -5.40735014e-02, -9.36852470e-02,  4.09044847e-02,  1.15336493e-01],
      dtype=float32),
      array([-0.05026853, -0.02421026, -0.03221055, -0.14009777,  0.02021943,
       -0.02261522, -0.08134355,  0.12336601,  0.07266331, -0.10568545,
        0.00092218, -0.03538591], dtype=float32)],
      dtype=object)

但我想要这样的输出：

array([[-0.02279311,  0.04733656,  0.02461601, ...,  0.02087441,
        -0.04001932,  0.00982925],
       [-0.03176997, -0.02204693,  0.01420259, ..., -0.07120648,
        -0.05186931, -0.04976927],
       [-0.01539093, -0.00272909, -0.02494676, ..., -0.02440629,
         0.01618153, -0.05321534],
       ...,
       [-0.05358865,  0.01567019, -0.03903598, ...,  0.01297285,
        -0.02813709, -0.08277859],
       [-0.00991264,  0.0100073 ,  0.04893894, ..., -0.03793963,
        -0.01426407, -0.04612683],
       [ 0.01175337, -0.02632044, -0.04074009, ..., -0.00123112,
         0.03152793, -0.00567225]], dtype=float32)

我该如何解决这种情况？因为使用数组输出我无法进行分类。

谢谢

Answer 1

不要使用数组构造函数来堆叠你的数组，但是：

vecs_train_LV = np.r_[[norm_vector(sentence, model_level, sw_indo) for sentence in data.data]]

或者：

vecs_train_LV = np.vstack([norm_vector(sentence, model_level, sw_indo) for sentence in data.data])