Table of Contents

First, new learning

1.1 Image input form of convolutional neural network

1.2 About compilation

1.3 class_names[np.argmax(pre[1])]

First, new learning

1.1 Image input form of convolutional neural network

The input of the convolutional neural network (CNN) is in the form of a tensor (image_height, image_width,
color_channels), which contains the image height, width and color information.There is no need to enter batch size.color_channels is (R, G, B) corresponding to the three color channels of RGB respectively.In this example, our CNN input, a picture from the fashion_mnist dataset, is of shape (28, 28, 1) i.e. a grayscale image.We need to assign the shape to the parameter input_shape when declaring the first layer.

model = models.Sequential([layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)), #convolution layer 1, convolution kernel 3*3layers.MaxPooling2D((2, 2)), #pooling layer 1, 2*2 samplinglayers.Conv2D(64, (3, 3), activation='relu'), #Convolution layer 2, convolution kernel 3*3layers.MaxPooling2D((2, 2)), #pooling layer 2, 2*2 samplinglayers.Conv2D(64, (3, 3), activation='relu'), #Convolution layer 3, convolution kernel 3*3layers.Flatten(), #Flatten layer, connecting the convolutional layer and the fully connected layerlayers.Dense(64, activation='relu'), #Full connection layer, further feature extractionlayers.Dense(10) #Output layer, output expected result])model.summary() # print network structure

So in the convolutional layer 1, the shape value of the image should be passed in

1.2 About compilation

Before you are ready to train your model, you need to set it up a bit more.The following is added in the compilation step of the model:
(1) loss function (loss): used to measure the accuracy of the model during training.You will want to minimize this function in order to "steer" the model in the right direction.

Loss functions include predicted value and actual squared difference (binary cross entropy), mean squared difference, etc.

(2) Optimizer ((optimizer): Determines how the model is updated based on the data it sees and its own loss function.

Help update parameters in real time

(3) metrics: used to monitor training and testing steps.The following examples use accuracy, which is the ratio of images that are correctly classified.

1.3 class_names[np.argmax(pre[1])]

See below

import numpy as npa = np.array([3, 1, 2, 4, 6, 1])b=np.argmax(a)# Take out the index corresponding to the maximum value of the element in a. At this time, the maximum value is 6, and the corresponding position index value is 4, (the index value starts from 0 by default)print(b)#4

Reference: np.argmax()_wanghua609's blog-CSDN blog_np.argmax

So np.argmax(pre[1]) is the index value i of the maximum confidence of the first image in the test set for the clothing in 10

From class_names[ i ]: take out the name