Question

我创建了一个类似于以下内容的ConvNet：

model = Sequential()
optimizer = Adam()

model.add(Lambda(lambda x: x / 127.5 - 1., input_shape=(28, 28, 1)))
model.add(Convolution2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(NUM_CLASSES, activation='softmax'))

model.compile(optimizer=optimizer, loss=keras.losses.categorical_crossentropy, metrics=['accuracy'])

我正在使用形状数据进行训练

X_train.shape = (48000, 28, 28, 1)
X_val.shape = (12000, 28, 28, 1)

而且效果很好。

但是，我现在想使用keras.evaluate()函数测试模型：

score = trained_model.evaluate(X_test, y_test, batch_size=128)
# X_test.shape = (10000, 28, 28, 1)
# y_test.shape (10000,)

导致以下错误：

ValueError: Error when checking target: expected dense_2 to have shape (10,) but got array with shape (1,)

鉴于我对训练，验证和测试集使用相同的形状，因此我不太理解此错误。

您介意解释我的错误是什么，以及如何解决该错误吗？

非常感谢！

编辑：输出trained_model.summary()

_________________________________________________________________
Layer (type)                 Output Shape              Param #
=================================================================
lambda_1 (Lambda)            (None, 28, 28, 1)         0
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 26, 26, 64)        640
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 13, 13, 64)        0
_________________________________________________________________
dropout_1 (Dropout)          (None, 13, 13, 64)        0
_________________________________________________________________
flatten_1 (Flatten)          (None, 10816)             0
_________________________________________________________________
dense_1 (Dense)              (None, 128)               1384576
_________________________________________________________________
dropout_2 (Dropout)          (None, 128)               0
_________________________________________________________________
dense_2 (Dense)              (None, 10)                1290
=================================================================
Total params: 1,386,506
Trainable params: 1,386,506
Non-trainable params: 0

评论中给出的解决方案
我忘记了一次热更新我的y_train，y_val和y_test数据。解决：

from keras.utils.np_utils import to_categorical
y_train = to_categorical(y_train)

Answer 1

该错误表明目标（y）的形状应为单热编码，每个样本包含10个元素。您证明y_test的形状为（10000，），不是一次性编码。

您可以执行以下操作：

y_test = kera.utils.np_utils.to_categorical(y_test)

Keras model.evaluate（）失败

1 个答案: