函数式API简介

函数式API简介

转自：https://www.cnblogs.com/miraclepbc/p/14312152.html

导入相关库以及数据加载

相关库导入：

import tensorflow as tf
from tensorflow import keras
import matplotlib.pyplot as plt
%matplotlib inline

数据加载：

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

数据归一化：

train_images = train_images / 255.0
test_images = test_images / 255.0

函数式定义模型

输入：

input = keras.Input(shape = (28, 28))

这里的意思就是可以传任意28*28的数据

模型定义：

x = keras.layers.Flatten()(input)
x = keras.layers.Dense(32, activation = 'relu')(x)
x = keras.layers.Dropout(0.5)(x)
x = keras.layers.Dense(64, activation = 'relu')(x)

输出：

output = keras.layers.Dense(10, activation = 'softmax')(x)

构建模型：

model = keras.Model(inputs = input, outputs = output)
model.summary()

模型编译

model.compile(
    optimizer = 'adam',
    loss      = 'sparse_categorical_crossentropy',
    metrics   = ['acc']
)

模型训练

history = model.fit(
    train_images,
    train_labels,
    epochs = 30,
    validation_data = (test_images, test_labels)
)