Keras 服饰图片分类1——数据调用，模型构建

从Keras库里调取数据，训练数据，验证数据，测试数据

import keras
fashion_mnist = keras.datasets.fashion_mnist
(x_train_all, y_train_all), (x_test, y_test) = fashion_mnist.load_data()
x_valid, x_train = x_train_all[:5000], x_train_all[5000:]
y_valid, y_train = y_train_all[:5000], y_train_all[5000:]

print(x_valid.shape, y_valid.shape)
print(x_train.shape, y_train.shape)
print(x_test.shape, y_test.shape)

输出结果：
(5000, 28, 28) (5000,)
(55000, 28, 28) (55000,)
(10000, 28, 28) (10000,)

调取一张图片

import matplotlib.pyplot as plt
def show_single_image(img_arr):
    plt.imshow(img_arr, cmap="binary")
    plt.show()

show_single_image(x_train[0])

cmap——颜色图谱

def show_imgs(n_rows, n_cols, x_data, y_data, class_names):
    assert len(x_data) == len(y_data)
    assert n_rows * n_cols < len(x_data)
    plt.figure(figsize = (n_cols * 1.4, n_rows * 1.6))
    num = n_rows * n_cols
    for i in range(num):
        index = i + 1
        plt.subplot(n_rows,n_cols,index)
        plt.imshow(x_train[index-1],cmap="binary",interpolation="nearest")
        plt.axis("off")
        plt.title(class_names[y_train[index-1]])
    plt.show()
show_images(3,4,x_train,y_train,class_names)

assert 断言函数，验证，条件限制，不符合条件就停止跳出；
figsize 画布尺寸宽*高（几列乘以几行）；
interpolation=“nearest” 如果显示分辨率与图像分辨率不同时，它将产生一个图像，其中像素显示为多个像素的正方形。

model = keras.models.Sequential()
model.add(keras.layers.Flatten(input_shape=[28,28]))
model.add(keras.layers.Dense(300,activation="relu"))
model.add(keras.layers.Dense(100,activation="relu"))
model.add(keras.layers.Dense(10,activation="sofemax"))

也可写成这样：

model = keras.models.Sequential([
    keras.layers.Flatten(input_shape=[28,28]),
    keras.layers.Dense(300,activation="relu"),
    keras.layers.Dense(100,activation="relu"),
    keras.layers.Dense(10,activation="softmax")
])

relu: y=max(0,x) , x>0, 取x, x<0，取0
sofemax 多分类输出，将向量变成概率分布，x = [x1,x2,x3]
y=[e^x1 /sum,e^x2 /sum,e^x3 /sum] sum = e^x1 +e^x2+ e^x3
y = [y1,y2,y3]， y1+y2+y3=1 , 0<y1,y2,y3<1

model.compile(loss="sparse_categorical_crossentropy",
              optimizer = "sgd",
              metrics = ["accuracy"])

sparse_categorical_crossentropy ——交叉熵函数 y的输入为index（一个数值）
categorical_crossentropy —— y的输入为one—hot([0,0,1])
optimizer梯度下降（sgd随机梯度下降，bgd批量梯度下降，msgd小批量梯度下降）
metrics准确率计算方法

accuracy	真实值为int，预测值也为int：[0, 1, 3, 3, 4, 2]
categorical_accuracy	真实值为onehot标签 ，预测值为向量值
binary_accuracy	1.二进制标签[0, 0, 0, 1, 1, 0]。 2.有threshold
sparse_categorical_accuracy	真实值为int， 预测值为向量值

print(model.layers)

查看模型构建框架

print(model.summary())

查看模型概况，都有多少参数进行训练 样本数[None,784]784维矩阵——经全连接变成300维矩阵
[None,784]w[784,300]+b[300]–[None,300] w权重，b偏执 784300+300=23550

history = model.fit(x_train, y_train, epochs=10,
                    validation_data=(x_valid, y_valid))

type(history)——查询类型
tensorflow.python.keras.callbacks.History
print(history.history)——查询loss,accuracy等数值变化

import pandas as pd
def plot_learning_curves(history):
    pd.DataFrame(history.history).plot(figsize=(8, 5))
    plt.grid(True)
    plt.gca().set_ylim(0, 1)
    plt.show()

plot_learning_curves(history)

grid(True) 显示网格线
plt.gca() 获取坐标轴信息
set_ylim(0, 1) 设置y轴限制，0-1区间

数据转化成one-hot结构
convert to 1-of-k encoding index变one-hot

num_classes = 10
y_train = keras.utils.to_categorical(y_train,num_classes)
y_test = keras.utils.to_categorical(y_test,num_classes)