tf环境下，VGG16训练mnist ，并达到一个很高的准确率，给出完整代码，打印运算结果，打印运行时间 并可视化

首先需要导入相关的库：

import tensorflow as tf
from tensorflow.keras.datasets import mnist
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Conv2D, MaxPooling2D, Flatten
from tensorflow.keras.utils import to_categorical
import numpy as np
import matplotlib.pyplot as plt
import time

接下来加载MNIST数据集并进行预处理：

# 加载数据集
(train_images, train_labels), (test_images, test_labels) = mnist.load_data()

# 将数据转换为float32类型
train_images = train_images.astype('float32')
test_images = test_images.astype('float32')

# 将像素值缩放到0-1之间
train_images /= 255
test_images /= 255

# 将标签转换为one-hot编码
train_labels = to_categorical(train_labels, num_classes=10)
test_labels = to_categorical(test_labels, num_classes=10)

然后构建VGG16模型：

model = Sequential()

# Block 1
model.add(Conv2D(64, (3, 3), activation='relu', padding='same', input_shape=(28, 28, 1)))
model.add(Conv2D(64, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))

# Block 2
model.add(Conv2D(128, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(128, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))

# Block 3
model.add(Conv2D(256, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(256, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(256, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))

# Block 4
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))

# Block 5
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(Conv2D(512, (3, 3), activation='relu', padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))

# 全连接层
model.add(Flatten())
model.add(Dense(4096, activation='relu'))
model.add(Dense(4096, activation='relu'))
model.add(Dense(10, activation='softmax'))

# 打印模型结构
model.summary()

接着进行模型训练：

# 编译模型
model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

# 记录开始时间
start_time = time.time()

# 训练模型
model.fit(train_images.reshape(-1, 28, 28, 1), train_labels, epochs=5, batch_size=64, 
          validation_data=(test_images.reshape(-1, 28, 28, 1), test_labels))

# 记录结束时间
end_time = time.time()

# 打印运行时间
print("运行时间：%.2f秒" % (end_time - start_time))

最后进行模型评估和可视化：

# 评估模型
test_loss, test_acc = model.evaluate(test_images.reshape(-1, 28, 28, 1), test_labels)
print('测试集准确率：%.2f%%' % (test_acc * 100))

# 可视化
predictions = model.predict(test_images.reshape(-1, 28, 28, 1))
fig, axes = plt.subplots(10, 10, figsize=(15, 15))
fig.subplots_adjust(hspace=0.3, wspace=0.3)
for i, ax in enumerate(axes.flat):
    ax.imshow(test_images[i], cmap='binary')
    ax.set_xticks([])
    ax.set_yticks([])
    ax.set_xlabel(np.argmax(predictions[i]))
plt.show()