使用 Keras Sequential() 编译时出现 ValueError

Question

我正在尝试使用 Keras 模块使用 Sequential() 编译数据集，但我得到一个值错误：

ValueError: Error when checking model input: expected dense_input_1 to have shape (None, 33) but got array with shape (32, 36)

我多次检查我的代码，但找不到任何可能的错误。

我有一个包含 32 个项目的数据集，所有项目都转换为浮点数。

这是我的神经网络的代码：

# Build neural network
# Sequential
model = Sequential()

# Neural network
model.add(Dense(36, input_dim=34, init='uniform', activation='sigmoid' ))
model.add(Dense(32, init='uniform', activation='sigmoid'))
model.add(Dense(32, init='uniform', activation='sigmoid'))
model.add(Dense(32, init='uniform', activation='sigmoid'))
model.add(Dense(33, init='uniform', activation='sigmoid'))

# Compile model
model.compile(loss='mean_squared_logarithmic_error', optimizer='SGD', metrics=['accuracy'])

# Fit model
history = model.fit(X, Y, nb_epoch=20, validation_split=0.2, batch_size=3)

这是我收到的完整错误消息：

Traceback (most recent call last):
  File "/Users/cliang/Desktop/Laurence/Python/Programs/Python/Collaborative_Projects/Cancer_screening/neural_network_alls_1.py", line 111, in <module>
    history = model.fit(X, Y, nb_epoch=20, validation_split=0.2, batch_size=3)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/keras/models.py", line 672, in fit
    initial_epoch=initial_epoch)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/keras/engine/training.py", line 1116, in fit
    batch_size=batch_size)
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/keras/engine/training.py", line 1029, in _standardize_user_data
    exception_prefix='model input')
  File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/keras/engine/training.py", line 124, in standardize_input_data
    str(array.shape))
ValueError: Error when checking model input: expected dense_input_1 to have shape (None, 34) but got array with shape (32, 36)

Answer 1

正如错误所说，您的输入数据形状和第一层之间存在不匹配。您清楚地定义您的输入维度是（特征数）是 34 input_dim=34，尽管您正在传递具有 36 个特征的数据。

我认为您混淆了隐藏层 36 的神经元数量和输入数据 34 的数量。要么从数据中删除两列，要么更改 input_dim=36。