神经网络

import tensorflow as tf
from tensorflow.keras.layers import Dense, Flatten
from tensorflow.keras.models import Sequential
from tensorflow.keras.losses import SparseCategoricalCrossentropy
from tensorflow.keras.optimizers import Adam
import numpy as np

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  # 归一化像素值

print(x_train.shape)
print(x_test.shape)

model = Sequential([
    Flatten(input_shape=(28, 28)),  # 将28x28的图像展平成784个元素的一维数组
    Dense(128, activation='relu'),  # 第一层隐藏层，128个节点，ReLU激活函数
    Dense(64, activation='relu'),   # 第二层隐藏层，64个节点，ReLU激活函数
    Dense(10, activation='softmax')  # 输出层，10个节点，对应10个类别，使用softmax激活函数进行多分类
])

model.compile(optimizer=Adam(),              # 使用Adam优化器
              loss=SparseCategoricalCrossentropy(),  # 使用稀疏分类交叉熵损失函数
              metrics=['accuracy'])                 # 评估指标为准确率

model.summary()
			  
model.fit(x_train, y_train, epochs=10, validation_data=(x_test, y_test))  # 训练模型并在测试集上验证

test_loss, test_acc = model.evaluate(x_test, y_test, verbose=2)
print('\nTest accuracy:', test_acc)

predictions = model.predict(x_test)
print("Predictions:\n", predictions)  # 查看预测结果（概率分布）

imagePath = "D:\\work\\mygithub\\docs\\intoAI\\src\\chapter2\\images\\intoAI_2_0.png"
img = tf.keras.utils.load_img(imagePath, color_mode='grayscale')
print(img.size)  # 输出图像尺寸
print(img.mode)  # 输出图像模式，应为 'L' 表示灰度图像
img_array = (np.expand_dims(img,0))
print(img_array.shape)
predictions = model.predict(img_array)
print(predictions)