arrays - 检查目标 : expected dense_3 to have shape (10, 时出错，但得到形状为 (2,) 的数组？即使标签是 one-hot 编码的

Question

我是机器学习和 Keras 的新手，我正在摆弄一些代码。我正在尝试制作一个图像分类器，可以确定图片是否是猫。 我的问题是，当我将 test_set_y 和 train_set_y 传递到 model.fit() 时，数组形状不匹配。

我在堆栈溢出中搜索了同样的问题，许多解决方案都包括对标签进行one-hot编码。然而，在我对标签进行一次性编码后，问题仍然存在。

def load_dataset():
    train_dataset = h5py.File('cat/train_catvnoncat.h5', "r")
    train_set_x_orig = np.array(train_dataset["train_set_x"][:]) # your train set features
    train_set_y_orig = np.array(train_dataset["train_set_y"][:]) # your train set labels

    test_dataset = h5py.File('cat/test_catvnoncat.h5', "r")
    test_set_x_orig = np.array(test_dataset["test_set_x"][:]) # your test set features
    test_set_y_orig = np.array(test_dataset["test_set_y"][:]) # your test set labels

    classes = np.array(test_dataset["list_classes"][:]) # the list of classes

    train_set_y_orig = train_set_y_orig.reshape((1, train_set_y_orig.shape[0]))
    test_set_y_orig = test_set_y_orig.reshape((1, test_set_y_orig.shape[0]))

    return train_set_x_orig, train_set_y_orig, test_set_x_orig, test_set_y_orig, classes

train_dataset = h5py.File('cat/train_catvnoncat.h5', "r")
test_dataset = h5py.File('cat/test_catvnoncat.h5', "r")
# Loading the data (cat/non-cat)
train_set_x_orig, train_set_y, test_set_x_orig, test_set_y, classes = load_dataset()

# Example of a picture
index = 78
example = train_set_x_orig[index]
plt.imshow(train_set_x_orig[index])
plt.show()
print("y = " + str(train_set_y[:, index]) + ", it's a '" + classes[np.squeeze(train_set_y[:, index])].decode("utf-8") + "' picture.")

print(train_set_x_orig.shape, test_set_x_orig.shape, train_set_y.shape, test_set_y.shape)

# One hot encode the labels------
train_set_y = to_categorical(train_set_y, num_classes=2)
test_set_y = to_categorical(test_set_y, num_classes=2)
print(train_set_y.shape, test_set_y.shape)

train_set_y = np.reshape(train_set_y, (209, 2))
test_set_y = np.reshape(test_set_y, (50, 2))

print(train_set_y.shape, test_set_y.shape)



# CNN ---------
# Forming model
model = Sequential()

# Adding layers
model.add(Conv2D(64, kernel_size=5, strides=1, padding="Same", activation="relu", input_shape=(64, 64, 3)))
model.add(MaxPooling2D(padding="same"))

model.add(Conv2D(128, kernel_size=5, strides=1, padding="same", activation="relu"))
model.add(MaxPooling2D(padding="same"))
model.add(Dropout(0.3))

model.add(Flatten())

model.add(Dense(256, activation="relu"))
model.add(Dropout(0.3))

model.add(Dense(512, activation="relu"))
model.add(Dropout(0.3))

model.add(Dense(10, activation="softmax"))

# Compiling the model 
model.compile(optimizer="adam", loss="categorical_crossentropy", metrics=["accuracy"])

# Training the model
model.fit(train_set_x_orig, train_set_y, batch_size=50, epochs=30, validation_data=(test_set_x_orig, test_set_y))

# Evaluate
train_loss_score = model.evaluate(train_set_x_orig, train_set_y)
test_loss_score = model.evaluate(test_set_x_orig, test_set_y)
print(train_loss_score)
print(test_loss_score)

我希望模型能够训练，最后给我损失和分数，但我得到“ValueError:检查目标时出错:预期dense_3具有形状(10，)但得到形状为(2，)的数组”

Answer 1

看看你的数据，你有两个类，你是 one_hot 编码的:

train_set_y = to_categorical(train_set_y, num_classes=2)

但是在您的模型中，您输出大小为 10 的张量:

model.add(Dense(10, activation="softmax"))

这不一致!

将最后一层更改为:

model.add(Dense(2, activation="softmax"))

它会起作用的!