Tensorflow Keras 形状不匹配

Question

在尝试实现许多教程用来向您介绍神经网络的标准 MNIST 数字识别器时，我遇到了错误

ValueError: Shape mismatch: The shape of labels (received (1,)) should equal the shape of logits except for the last dimension (received (28, 10)).

我想用 from_tensor_slices处理数据，因为我想将代码应用于数据来自 CSV 文件的另一个问题。无论如何，这是在行 model.fit(...) 中产生错误的代码

import tensorflow as tf

train_dataset, test_dataset = tf.keras.datasets.mnist.load_data()
train_images, train_labels = train_dataset
train_images = train_images/255.0
train_dataset_tensor = tf.data.Dataset.from_tensor_slices((train_images, train_labels))

num_of_validation_data = 10000
validation_data = train_dataset_tensor.take(num_of_validation_data)
train_data = train_dataset_tensor.skip(num_of_validation_data)

model = tf.keras.Sequential([
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(100, activation='sigmoid'),
    tf.keras.layers.Dense(10, activation='softmax')
])

model.compile(
    optimizer=tf.keras.optimizers.Adam(learning_rate=1e-3), 
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    metrics=['accuracy']
)

model.fit(train_data, batch_size=50, epochs=5)

performance = model.evaluate(validation_data)

我不明白形状在哪里 (28, 10)的 logits 来自，我以为我正在展平图像，基本上是从 2D 图像中制作一个 1D 矢量？我怎样才能防止错误？

Answer 1

您可以使用以下代码

mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

x_train = x_train[..., tf.newaxis]
x_test = x_test[..., tf.newaxis]

train_ds = tf.data.Dataset.from_tensor_slices(
    (x_train, y_train)).shuffle(10000).batch(32)
test_ds = tf.data.Dataset.from_tensor_slices((x_test, y_test)).batch(32)

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(100, activation='sigmoid'),
    tf.keras.layers.Dense(10, activation='softmax')
])

model.compile(
    optimizer=tf.keras.optimizers.Adam(learning_rate=1e-3), 
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    metrics=['accuracy']
)

model.fit(train_ds)