LeNet-5训练神经网络训练

LeNet-5训练

导包

import tensorflow as tf 

from tensorflow.keras import layers, models, datasets, optimizers 

加载Fashion-MNIST数据集 

(train_images, train_labels), (test_images, test_labels) = datasets.fashion_mnist.load_data() 

归一化像素值到[0, 1]区间 

train_images = train_images.astype("float32") / 255 

test_images = test_images.astype("float32") / 255

对标签进行分类编码 

train_labels = tf.keras.utils.to_categorical(train_labels, 10) 

test_labels = tf.keras.utils.to_categorical(test_labels, 10)

定义LeNet-5模型

model = models.Sequential([ 

    layers.Conv2D(6, (5, 5), activation='relu', input_shape=(28, 28, 1)), 

    layers.AveragePooling2D((2, 2)), 

    layers.Conv2D(16, (5, 5), activation='relu'), 

    layers.AveragePooling2D((2, 2)), 

    layers.Conv2D(120, (3, 3), activation='relu', padding='valid'), 

# 注意：这里可能需要调整以避免过拟合或尺寸问题 

    layers.Flatten(), 

    layers.Dense(84, activation='relu'), 

    layers.Dense(10, activation='softmax') 

]) 

编译模型 

model.compile(optimizer=optimizers.Adam(), 

              loss='categorical_crossentropy', 

              metrics=['accuracy']) 

训练模型 

model.fit(train_images, train_labels, epochs=10, batch_size=64, validation_data=(test_images, test_labels))

评估模型

test_loss, test_acc = model.evaluate(test_images, test_labels) 

print(f'Test Accuracy: {test_acc:.4f}')

训练模型保存

save_path = r'D:\\图像处理、深度学习\\训练保存\\LeNet-5.h5'

model.save(save_path)