tensorflow 學習筆記 2

Github上有個學習tensorflow的git

https://github.com/nfmcclure/tensorflow_cookbook

內容完整，買了中文版，但是卻很多細節，所以還是直接看github原版好了。電子書是ipynb格式的，ubuntu看起來很容易。

從第一個example開始：

電子書

https://github.com/nfmcclure/tensorflow_cookbook/tree/master/01_Introduction/02_Creating_and_Using_Tensors

代碼鏈接：

https://github.com/nfmcclure/tensorflow_cookbook/blob/master/01_Introduction/02_Creating_and_Using_Tensors/02_tensors.py

代碼比較短，貼一下好了，對照相關部分的電子書和tensorboard的輸出還是能加強一些理解的

# Tensors
	#----------------------------------
	#
	# This function introduces various ways to create
	# tensors in TensorFlow
	import tensorflow as tf
	from tensorflow.python.frameworkimport ops
	ops.reset_default_graph()
	# Introduce tensors in tf
	# Get graph handle
	sess = tf.Session()
	my_tensor = tf.zeros([1,20])
	# Declare a variable
	my_var = tf.Variable(tf.zeros([1,20]))
	# Different kinds of variables
	row_dim = 2
	col_dim = 3
	# Zero initialized variable
	zero_var = tf.Variable(tf.zeros([row_dim, col_dim]))
	# One initialized variable
	ones_var = tf.Variable(tf.ones([row_dim, col_dim]))
	# shaped like other variable
	sess.run(zero_var.initializer)
	sess.run(ones_var.initializer)
	zero_similar = tf.Variable(tf.zeros_like(zero_var))
	ones_similar = tf.Variable(tf.ones_like(ones_var))
	sess.run(ones_similar.initializer)
	sess.run(zero_similar.initializer)
	# Fill shape with a constant
	fill_var = tf.Variable(tf.fill([row_dim, col_dim],-1))
	# Create a variable from a constant
	const_var = tf.Variable(tf.constant([8,6, 7, 5, 3, 0, 9]))
	# This can also be used to fill an array:
	const_fill_var = tf.Variable(tf.constant(-1,shape=[row_dim, col_dim]))
	# Sequence generation
	linear_var = tf.Variable(tf.linspace(start=0.0,stop=1.0,num=3))# Generates [0.0, 0.5, 1.0] includes the end
	sequence_var = tf.Variable(tf.range(start=6,limit=15,delta=3))# Generates [6, 9, 12] doesn't include the end
	# Random Numbers
	# Random Normal
	rnorm_var = tf.random_normal([row_dim, col_dim],mean=0.0,stddev=1.0)
	# Add summaries to tensorboard
	merged = tf.summary.merge_all()
	# Initialize graph writer:
	writer = tf.summary.FileWriter("/tmp/variable_logs",graph=sess.graph)
	# Initialize operation
	initialize_op = tf.global_variables_initializer()
	# Run initialization of variable
	sess.run(initialize_op)

在貼個tensorboard的graph：

如果點擊一個具體的變量，會看到與語句的關係，比如

tf.Variable(tf.zeros([1,20]))

對應的就是：

所以很簡單的一個理解就是tensorflow框架對tensor做了運算，然後通過tensorboard可以直觀的觀察到。

一些額外的學習資料：

# Additional Resources

###Official Resources:

 - [TensorFlow Python API](https://www.tensorflow.org/api_docs/python/)
 - [TensorFlow on Github](https://github.com/tensorflow/tensorflow)
 - [TensorFlow Tutorials](https://www.tensorflow.org/tutorials/)
 - [Udacity Deep Learning Class](https://www.udacity.com/course/deep-learning--ud730)
 - [TensorFlow Playground](http://playground.tensorflow.org/)

###Github Tutorials and Examples:

 - [Tutorials by pkmital](https://github.com/pkmital/tensorflow_tutorials)
 - [Tutorials by nlintz](https://github.com/nlintz/TensorFlow-Tutorials)
 - [Examples by americdamien](https://github.com/aymericdamien/TensorFlow-Examples)
 - [TensorFlow Workshop by amygdala](https://github.com/amygdala/tensorflow-workshop)

###Deep Learning Resources

 - [Efficient Back Prop by Yann LeCun, et. al.](http://yann.lecun.com/exdb/publis/pdf/lecun-98b.pdf)
 - [Online Deep Learning Book, MIT Press](http://www.deeplearningbook.org/)
 - [An Overview of Gradient Descent Algorithms by Sebastian Ruder](http://sebastianruder.com/optimizing-gradient-descent/)
 - [Stochastic Optimization by John Duchi, et. al.](http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf)
 - [ADADELTA Method by Matthew Zeiler](http://arxiv.org/abs/1212.5701)
 - [A Friendly Introduction to Cross-Entropy Loss by Rob DiPietro](http://rdipietro.github.io/friendly-intro-to-cross-entropy-loss/)


###Additional Resources

 - [A Curated List of Dedicated TensorFlow Resources](https://github.com/jtoy/awesome-tensorflow/)

###Arxiv Papers

 - [TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems](http://arxiv.org/abs/1603.04467)
 - [TensorFlow: A system for large-scale machine learning](http://arxiv.org/abs/1605.08695)
 - [Distributed TensorFlow with MPI](https://arxiv.org/abs/1603.02339)
 - [Comparative Study of Deep Learning Software Frameworks](https://arxiv.org/abs/1511.06435)
 - [Wide & Deep Learning for Recommender Systems](https://arxiv.org/abs/1606.07792)