[Python] CNN

昨天用了一般的NN来做影像分类，但其实同样的情况用CNN会有效率很多，今天就来建立CNN再应用一次。

下载dataset和packages

import tensorflow as tf
from tensorflow import keras
fashion_mnist = keras.datasets.fashion_mnist
(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()
X_train = (X_train / 255.0).reshape(60000, 28, 28, 1)
X_test = (X_test / 255.0).reshape(10000, 28, 28, 1)
class_names = ["T-shirt/top", "Trouser", "Pullover", "Dress", "Coat",
                   "Sandal", "Shirt", "Sneaker", "Bag", "Ankle boot"]

开始建模，记得第一层layer要放input shape，layer中重要的参数有：

1. filters:每层中filter的数量
2. kernal_size:filter的宽度与长度，必须要是奇数
3. strides: 步伐大小
4. padding:是不是要处理边界，一般来说设定都是 “same”，也就是要处理

model = keras.models.Sequential()
model.add(keras.layers.Conv2D(filters=7, kernel_size=5, strides=1, padding="same", activation="relu",input_shape=[28, 28, 1]))
model.add(keras.layers.MaxPool2D(pool_size=2))

model.add(keras.layers.Conv2D(filters=7, kernel_size=3, strides=1, padding="same", activation="relu"))
model.add(keras.layers.MaxPool2D(pool_size=2))

model.add(keras.layers.Flatten())
model.add(keras.layers.Dense(50, activation="relu"))
model.add(keras.layers.Dense(50, activation="relu"))
model.add(keras.layers.Dropout(0.3))
model.add(keras.layers.Dense(10, activation="softmax"))


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

history = model.fit(X_train, y_train, epochs=10, validation_split=0.1)
model.evaluate(X_test, y_test)

[reference]
https://blog.csdn.net/zz2230633069/article/details/88544747
https://www.geeksforgeeks.org/keras-conv2d-class/
https://www.cnblogs.com/yjybupt/p/11646846.html
https://woj.app/6491.html