tensorflow uses flags to define command line parameters

This article mainly introduces how tensorflow uses flags to define command line parameters. Now I share it with you and give it as a reference. Let’s take a look together

tf defines tf.app.flags to support accepting command line parameters, which is equivalent to accepting argv.

import tensorflow as tf

#第一个是参数名称，第二个参数是默认值，第三个是参数描述
tf.app.flags.DEFINE_string('str_name', 'def_v_1',"descrip1")
tf.app.flags.DEFINE_integer('int_name', 10,"descript2")
tf.app.flags.DEFINE_boolean('bool_name', False, "descript3")

FLAGS = tf.app.flags.FLAGS

#必须带参数，否则：'TypeError: main() takes no arguments (1 given)';  main的参数名随意定义，无要求
def main(_): 
  print(FLAGS.str_name)
  print(FLAGS.int_name)
  print(FLAGS.bool_name)

if __name__ == '__main__':
  tf.app.run() #执行main函数

Execution:

[root@AliHPC-G41-211 test]# python tt.py
def_v_1
10
False
[root@AliHPC-G41-211 test]# python tt.py --str_name test_str --int_name 99 --bool_name True
test_str
99
True

Related recommendations:

How to export TensorFlow’s model network to a single file

The above is the detailed content of tensorflow uses flags to define command line parameters. For more information, please follow other related articles on the PHP Chinese website!