我有一个CNN模型,可以将word2vec矩阵作为输入正常运行。现在,我正在尝试使用tf-idf功能作为CNN模型的输入。我的模型有2个卷积层。
vectorizer = TfidfVectorizer(max_features=10000, use_idf=True)
vectorizer = vectorizer.fit(train_sentences)
tf_len = len(vectorizer.vocabulary_)
TF_X_train = vectorizer.transform(train_sentences).astype('float64')
TF_X_test = vectorizer.transform(test_sentences).astype('float64')
TF_X_val = vectorizer.transform(val_sentences).astype('float64')
input = Input(shape=(tf_len,1))
drop20 = SpatialDropout1D(0.3)(input)
conv2 = Conv1D(filters=128, kernel_size=5, activation='relu')(drop20)
drop21 = Dropout(0.5)(conv2)
conv22 = Conv1D(filters=64, kernel_size=5, activation='relu')(drop21)
drop22 = Dropout(0.5)(conv22)
pool2 = MaxPooling1D(pool_size=2)(drop22)
flat2 = Flatten()(pool2)
out = Dense(8, activation='sigmoid')(flat2)
model = Model(input, out)
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
model.summary()
我收到以下错误。请提供任何提示来解决我的问题。
我也尝试将输入层更改为Input(batch_shape=(None,tf_len, 1))
,但遇到相同的错误。
[ValueError:检查输入时出错:预期input_1具有3个维度,但数组的形状为(1000,5008)
即使在注释部分中已经存在,也请在此(答案)部分中指定答案。
添加下面的代码行,
TF_X_train.reshape(TF_X_train.shape[0], TF_X_train.shape[1],1)
并将输入层更改为
Input(batch_shape=(None, tf_len, 1))
已修复错误。