from tensorflow.examples.tutorials.mnist import input_data改进代码

The `input_data` module has been deprecated since TensorFlow 1.14. You can use the `tf.keras.datasets` module to load the MNIST dataset instead. Here's an example code:

import tensorflow as tf
from tensorflow import keras

# Load MNIST dataset
(x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data()

# Normalize pixel values
x_train = x_train / 255.0
x_test = x_test / 255.0

# Define the model
model = keras.Sequential([
    keras.layers.Flatten(input_shape=(28, 28)),
    keras.layers.Dense(128, activation='relu'),
    keras.layers.Dense(10, activation='softmax')
])

# Compile the model
model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

# Train the model
model.fit(x_train, y_train, epochs=5)

# Evaluate the model
test_loss, test_acc = model.evaluate(x_test, y_test)
print('Test accuracy:', test_acc)

This code loads the MNIST dataset using `keras.datasets.mnist.load_data()`, normalizes the pixel values to the range [0, 1], defines a simple neural network with two dense layers, compiles the model using the Adam optimizer and sparse categorical cross-entropy loss, trains the model for 5 epochs, and evaluates the model on the test set.