Deep learning model code basic structure

if __name__=='__main__':

 

# args 선언

(dataset)

-dataset

-dataroot

 

(model)

-model 종류

 

(optimization)

-batch_size

-learning_rate

-step_size

 

(loss)

-loss_function

 

(options)

-seed

-the number of gpu

-use gpu

-eval mode

 

#main

(seed 선언)

(gpu 사용)

(dataset)

(model 선언)

- model

- criterion

 

 

(train)

-train

() load_network

-val

-save_network

 

result save & printing

tensorboard

git