Using keras in tensorflow to build convolutional neural network

Application tensorflow Medium keras Build convolutional neural network

Six steps

1、 Import tensorflow library

import tensorflow.keras as keras

2、 Prepare training data

x_train,y_train
x_test,y_test

3、 Build a network structure

• Generate a container to store the network structure

model = keras.models.Sequential() # Describe each layer of network 

• Convolution layer

keras.layers.Conv2D(
	filters,  # Number of convolution kernels 
    kernel_size,  # Convolution kernel size   It's usually (3,3)
    strides=(1, 1),  # Sliding step   Default (1,1)
    padding='valid',  # Zero filling strategy  'valid' perhaps 'same'
    activation=None,  # Activation function   Commonly used relu,softmax,selu
    input_shape  # Define the input data style  (64,64,3)64*64 Three dimensional view of 
)

• Pooling layer

keras.layers.MaxPooling2D(
    pool_size=(2, 2),  # Pool layer size 
    strides=None,  # step 
    padding='valid',  # Zero filling strategy  'valid' perhaps 'same'
    data_format=None
)

• Flatten （ Change the data into one dimension , It is often used in the transition from convolution layer to full connection layer ）

keras.layers.Flatten()

• Fully connected layer

keras.layers.Dense(
    units,  #  Dimension of output space 
    activation=None, #  Activation function   Commonly used relu,softmax,selu
    use_bias=True, #  Boolean value , Whether to use offset vector 
)

• Dropout layer （ Prevent too fitting , Improve the generalization ability of the model ）

keras.layers.Dropout(
	rate, #0-1 Decimal between   Percentage discarded 
    noise_shape=None, 
    seed=None # Random seeds 
)

4、 Print network structure and parameter statistics

model.summary()
summary Function is used to print network structure and parameter statistics

5、 Configure the optimizer for training 、 Loss function and accuracy evaluation criteria

model.compile(
    optimizer, # Optimizer 
    loss, #  Loss function 
    metrics # Network evaluation index 
)

• optimizer Parameters can be character string The optimizer name given by the form , It can also be in the form of a function , The learning rate can be set in the form of a function 、 Momentum and hyperparameters
  • “sgd” perhaps keras.optimizers.SGD(lr = Learning rate ,decay = Learning decay rate ,momentum = Momentum parameter ）
  • "adagrad’" perhaps keras.optimizers.Adagrad(lr = Learning rate , decay = Learning rate decay rate ）
  • "adadelta" perhaps keras.optimizers.Adadelta(lr = Learning rate , decay = Learning rate decay rate ）
  • "adam" perhaps keras.optimizers.Adam(lr = Learning rate ,decay = Learning rate decay rate ）
• lose Parameters can be given in string form Loss function Name , It can also be in the form of a function 、
  • "mse" perhaps keras.losses.MeanSquaredError()
  • “sparse_categorical_crossentropy” perhaps keras.losses.SparseCatagoricalCrossentropy(from_logits = False)
• Metrics Label network evaluation index
  • “accuracy” : y_ and y It's all numbers , Such as y_ = [1] y = [1] #y_ For real value ,y For the predicted value
  • "sparse_accuracy"y_ and y It's all based on a single hot code And probability distribution , Such as y_ = [0, 1, 0], y = [0.256, 0.695, 0.048]
  • "sparse_categorical_accuracy"y_ It is given in numerical form ,y In order to The unique heat code gives , Such as y_ = [1], y = [0.256 0.695, 0.048]

6、fit

model.fit( 
    x, y, 
    batch_size=32, 
    epochs=10, 
    verbose=1, 
    callbacks=None,
	validation_split=0.0, 
    validation_data=None, 
    shuffle=True, 
    class_weight=None, 
    sample_weight=None, 
    initial_epoch=0
)

• x： input data . If the model has only one input , that x The type is numpy
  array, If the model has multiple inputs , that x The type should be list,list Is corresponding to each input numpy array
• y： label ,numpy array
• batch_size： Integers , Specifies the gradient descent for each batch Number of samples included . One for training batch The sample will be calculated as a gradient descent , Optimize the target function one step .
• epochs： Integers , At the end of training epoch value , The training will be at the end of the day epoch Value , When there is no setting initial_epoch when , It is the total number of rounds of training , Otherwise, the total number of rounds of training is epochs - inital_epoch
• verbose： The log shows ,0 Output log information for non-standard output stream ,1 Record for the output progress bar ,2 For each epoch Output line record
• callbacks：list, The elements are keras.callbacks.Callback The object of . This list The callback function will be called at the appropriate time during the training , Refer to the callback function
• validation_split：0~1 The floating point number between , A percentage of the data used to specify the training set is used as the validation set . Validation sets will not be trained , And in each epoch End - of - test model metrics , Like the loss function 、 Precision etc. . Be careful ,validation_split The division of the shuffle Before , So if your data itself is ordered , You need to manually scramble it before you specify it validation_split, Otherwise, an uneven sample of the validation set may occur .
• validation_data： In the form of （X,y） Of tuple, Is the specified validation set . This parameter overrides validation_spilt.
• shuffle： Boolean or string , Is generally a Boolean value , Indicates whether the sequence of input samples is randomly scrambled during training . If string “batch”, It's used to deal with HDF5 The special case of data , It will be batch Internally scrambles the data .
• class_weight： Dictionaries , Mapping different categories to different weights , This parameter is used to adjust the loss function during training （ Only for training ）
• sample_weight： Weights of numpy
  array, Used to adjust the loss function during training （ For training purposes only ）. You can pass a 1D The vector with the same length as the sample is used to carry on the sample 1 Yes 1 A weighted , Or in the case of temporal data , The form of passing one is （samples,sequence_length） To assign different weights to the samples on each time step . In this case be sure to add when compiling the model sample_weight_mode=’temporal’.
• initial_epoch: Specified from this parameter epoch Start training , It's useful to continue the previous training .

fit The function returns a History The object of , Its History.history Attribute records the value of loss function and other indicators epoch Changing circumstances , If there is a verification set , Also contains the change of these indicators of the verification set