concept

some tf function

• 放图， 用法：name，tensor，max_outputs

tf.summary.image("asd",tensor,10)

• 统计直方图

tf.summary.histogram("asd",tensor)

• scalar 进行统计汇总，单个数值保存用 tf.summary.scalar，一般用来保存loss，accuary，学习率等数据，比较常用。

tf.summary.scalar("loss || accuracy",loss || accuracy)

• 定义了太多的summary汇总操作，逐一执行太麻烦，使用merge_all

merged = tf.summary.merge_all()

• train 和 test 分开写writer

train_writer = tf.summary.FileWriter(path,sess.graph)
test_writer = tf.summary.FileWriter(path,sess.graph)

• utility function

def variable_summaries(var):
    """Attach a lot of summaries to a Tensor (for TensorBoard visualization)."""
    with tf.name_scope('summaries'):
      mean = tf.reduce_mean(var)
      tf.summary.scalar('mean', mean)
      with tf.name_scope('stddev'):
        stddev = tf.sqrt(tf.reduce_mean(tf.square(var - mean)))
      tf.summary.scalar('stddev', stddev)
      tf.summary.scalar('max', tf.reduce_max(var))
      tf.summary.scalar('min', tf.reduce_min(var))
      tf.summary.histogram('histogram', var)

两种写法

1.建模型的时候使用name_scope，graph才会清楚一点，

tf.summary.scalar(tags, values)
tf.summary.histogram("name",tensor)
tf.summary.image("name",tensor,10)
# ...
merged = tf.summary.merge_all()
Writer = tf.summary.FileWriter(path, graph=sess.graph)

for epoch in XXX:
	run_options = tf.RunOptions(trace_level=tf.RunOptions.FULL_TRACE)
	run_metadata = tf.RunMetadata()
	summary = sess.run([merged,optimizer],options=run_options,
							run_metadata=run_metadata)
							
	Writer.add_run_metadata(run_metadata,"epoch %d"%epoch)
	Writer.add_summary(summary, global_step)

2.边写边 addsummary

summary_writer = tf.summary.FileWriter(logdir)
"""
.
.
.
"""
summary = tf.Summary(value=[
    tf.Summary.Value(tag="summary_tag", simple_value=0), 
    tf.Summary.Value(tag="summary_tag2", simple_value=1),
])
# x代表横轴坐标
summary_writer.add_summary(summary, x)

或者

summary_writer = tf.summary.FileWriter(logdir)
summary = tf.Summary()
summary.value.add(tag="summary_tag", simple_value=0)
summary.value.add(tag="summary_tag2", simple_value=1)
# x代表横轴坐标
summary_writer.add_summary(summary, x)

implement

demo1，第一种做法， 使用我的前一篇文章实作

code

demo2, 第二种做法

import tensorflow as tf
summary_writer = tf.summary.FileWriter('/tmp/test')
summary = tf.Summary(value=[
    tf.Summary.Value(tag="summary_tag", simple_value=0), 
    tf.Summary.Value(tag="summary_tag2", simple_value=1),
])
summary_writer.add_summary(summary, 1)

summary = tf.Summary(value=[
    tf.Summary.Value(tag="summary_tag", simple_value=1), 
    tf.Summary.Value(tag="summary_tag2", simple_value=3),
])
summary_writer.add_summary(summary, 2)

summary_writer.close()


--------------------- 
作者：EncodeTS 
来源：CSDN 
原文：https://blog.csdn.net/EncodeTS/article/details/54172807 
版权声明：本文为博主原创文章，转载请附上博文链接！

references:
https://blog.csdn.net/EncodeTS/article/details/54172807