Question

我正在喀拉拉邦写一篇LSTM，可以检测评论的毒性。我在X上训练模型。我在X上所做的操作

1。

max_features = 2000
tokenizer = Tokenizer(num_words=max_features)
tokenizer.fit_on_texts(data.values)

2。

dictionary = tokenizer.word_index

3。

with open('wordindex.json', 'w') as dictionary_file:
  json.dump(dictionary , dictionary_file)

4。

X = tokenizer.texts_to_sequences(data.values)
X = pad_sequences(X)

最终X的形状为(1396,)。我在这些课程上训练了模型

list_classes = ["toxic", "severe_toxic", "obscene", "threat", "insult", "identity_hate"]
y = train[list_classes].values

模型经过训练后，我尝试在输入中对其进行测试。我为转换输入pred = 'f you'而执行的步骤。

dictionary = json.load(open('wordindex.json'))
def convert_text_to_index_array(text):
         # one really important thing that `text_to_word_sequence` does
        # is make all texts the same length -- in this case, the length
        # of the longest text in the set.
        wordvec=[]
        for word in kpt.text_to_word_sequence(text) :
            if word in dictionary:
                wordvec.append([dictionary[word]])
            else:
                wordvec.append([0])

        return wordvec


   pred=convert_text_to_index_array(pred)

当我print(model.predict(pred))时。我得到一个输出ValueError: Error when checking model input: the list of Numpy arrays that you are passing to your model is not the size the model expected. Expected to see 1 array(s), but instead got the following list of 2 arrays: [array([[129]]), array([[6]])]... 当我尝试查看掠食者的形状时，出现错误：

    print(pred.shape)
AttributeError: 'list' object has no attribute 'shape'

我真的很沮丧，因为看起来我的pred是一维的。 array（[[129]]），array（[[6]]）]是我在json文件中编码的单词，但是为什么它们在单独的数组中？

检查模型输入时出错：传递给模型Keras的Numpy数组的列表

0 个答案: