tensorflow_api

tf.train.MonitoredTrainingSession

https://blog.csdn.net/MrR1ght/article/details/81006343

session

https://zhuanlan.zhihu.com/p/32869210

embedding look up

https://www.zhihu.com/question/52250059

estimator

https://zhuanlan.zhihu.com/p/74857888

 

globle step

global_step/sec：一种性能指标，显示我们在特定批处理（x 轴）中每秒处理了多少批处理（y 轴，梯度更新）

每秒扫了几个batch

mask用法

import tensorflow as tf
import numpy as np
sess=tf.Session()
 
a=np.array([[1,0,0],[0,1,1]])
a1=np.array([[3,2,3],[4,5,6]])

a=np.array([[3,3,4,4,4,4]])
b1=tf.ones_like(a)
b2=tf.zeros_like(a)
# a1=np.array([[3,2,3],[4,5,6]])
 
print(sess.run(tf.equal(a,3)))
print(sess.run(tf.where(tf.equal(a,3),a,b1)))
print(sess.run(tf.where(tf.equal(a,3),a,b2)))

分布式训练调参

https://developer.aliyun.com/article/705132

hook

https://gist.github.com/saravanabalagi/d37a139e3da48a77eb9f35de4176d653

优化器

http://blog.720ui.com/2018/tensorflow_03_dl_tensorflow_optimizer/#tf-train-AdagradOptimizer