我正在尝试使用以下ANN模型运行多类分类：
classifier = Sequential()
classifier.add(Dense(units = 9, kernel_initializer = 'uniform', activation = 'relu', input_dim = 18))
classifier.add(Dense(units = 9, kernel_initializer = 'uniform', activation = 'relu'))
classifier.add(Dense(units = 9, kernel_initializer = 'uniform', activation = 'relu'))
classifier.add(Dense(units = 6 ,kernel_initializer = 'uniform', activation = 'softmax'))
classifier.compile(optimizer = 'adam', loss = 'sparse_categorical_crossentropy', metrics = ['accuracy'])
classifier.fit(X_train, y_train, batch_size = 10, epochs = 100)
y_pred = classifier.predict(X_test)
X_train的格式是：
[[31 8 27 ... 2 7 5]
[31 8 11 ... 1 9 3]
[6 0 4 ... 1 9 3]
...
[55 55 134 ... 5 5 6]
[41 9 111 ... 1 3 0]
[19 9 28 ... 3 0 0]]
和y_train是：
[[0. 0. 0. 1. 0. 0.]
[0. 0. 0. 0. 1. 0.]
[0. 0. 0. 0. 1. 0.]
...
[0. 0. 0. 0. 0. 1.]
[0. 0. 0. 0. 0. 1.]
[0. 0. 0. 0. 0. 1.]]
X_train的形状是（352,18），y_train的形状是（352,6），X_test的形状是（152,18） .
当它运行时，它会给出以下错误：
Traceback (most recent call last):
File "H:\p36564\Project ZS\tst1.py", line 110, in <module>
classifier.fit(X_train, y_train, batch_size = 10, epochs = 100)
File "H:\p36564\lib\site-packages\keras\engine\training.py", line 950, in fit
batch_size=batch_size)
File "H:\p36564\lib\site-packages\keras\engine\training.py", line 787, in _standardize_user_data
exception_prefix='target')
File "H:\p36564\lib\site-packages\keras\engine\training_utils.py", line 137, in standardize_input_data
str(data_shape))
ValueError: Error when checking target: expected dense_3 to have shape (1,) but got array with shape (6,)
可能出现此错误的原因是什么？任何帮助，将不胜感激 .
1 回答
如果您提供了
y_train形状，请使用categorical_crossentropy作为丢失函数而不是sparse_categorical_crossentropy. 您的y_train是单热编码的，不是稀疏编码的 . 在您的情况下，稀疏编码将是一个如下所示的数组：要自己尝试一下，将
y_train转换为稀疏编码，如下所示：这将与
sparse_categorical_crossentropy一起用作损失函数（无需更改模型体系结构！）