tensorflow_cookbook：Ch 1: Getting Started with TensorFlow_(3,4)

Variables and Placeholders

Here we initialize a simple placeholder for a piece of data. We then feed a random number into an identity function. If viewed in Tensorboard, the following graph will be created.

#03_placeholders.py
# Placeholders
#----------------------------------
#
# This function introduces how to 
# use placeholders in Tensorflow

import numpy as np
import tensorflow as tf
from tensorflow.python.framework import ops
ops.reset_default_graph()

# Using Placeholders
sess = tf.Session()

x = tf.placeholder(tf.float32, shape=(4, 4))
y = tf.identity(x)

rand_array = np.random.rand(4, 4)
####
#merged = tf.merge_all_summaries()
merged = tf.summary.merge_all()
#writer = tf.train.SummaryWriter("/tmp/variable_logs", sess.graph_def)
writer = tf.summary.FileWriter("/tmp/variable_logs", sess.graph)
####
writer.close()
print(sess.run(y, feed_dict={x: rand_array}))

原代码会报错

AttributeError: module 'tensorflow' has no attribute 'merge_all_summaries'
AttributeError: module 'tensorflow.python.training.training' has no attribute 'SummaryWriter'

修改后结果

[[ 0.66682321  0.3293488   0.5245204   0.53379935]
 [ 0.07488234  0.10959593  0.68837297  0.17703928]
 [ 0.35069257  0.77648026  0.27833292  0.44115242]
 [ 0.76638317  0.15466152  0.22070079  0.94068062]]

Working with Matrices

Placeholder for future purposes.

# 04_matrices.py
#Matrices and Matrix Operations
#----------------------------------
#
# This function introduces various ways to create
# matrices and how to use them in Tensorflow

import numpy as np
import tensorflow as tf
from tensorflow.python.framework import ops
ops.reset_default_graph()

# Declaring matrices
sess = tf.Session()

# Declaring matrices

# Identity matrix
identity_matrix = tf.diag([1.0,1.0,1.0])
print(sess.run(identity_matrix))

# 2x3 random norm matrix
A = tf.truncated_normal([2,3])
print(sess.run(A))

# 2x3 constant matrix
B = tf.fill([2,3], 5.0)
print(sess.run(B))

# 3x2 random uniform matrix
C = tf.random_uniform([3,2])
print(sess.run(C))
print(sess.run(C)) # Note that we are reinitializing, hence the new random variabels

# Create matrix from np array
D = tf.convert_to_tensor(np.array([[1., 2., 3.], [-3., -7., -1.], [0., 5., -2.]]))
print(sess.run(D))

# Matrix addition/subtraction
print(sess.run(A+B))
print(sess.run(B-B))

# Matrix Multiplication
print(sess.run(tf.matmul(B, identity_matrix)))

# Matrix Transpose
print(sess.run(tf.transpose(C))) # Again, new random variables

# Matrix Determinant
print(sess.run(tf.matrix_determinant(D)))

# Matrix Inverse
print(sess.run(tf.matrix_inverse(D)))

# Cholesky Decomposition
print(sess.run(tf.cholesky(identity_matrix)))

# Eigenvalues and Eigenvectors
print(sess.run(tf.self_adjoint_eig(D)))

[[ 1.  0.  0.]
 [ 0.  1.  0.]
 [ 0.  0.  1.]]
[[ 0.4898375  -0.65810621  0.20719798]
 [-0.45387158 -0.72594666  0.2478656 ]]
[[ 5.  5.  5.]
 [ 5.  5.  5.]]
[[ 0.37838769  0.72760344]
 [ 0.2227149   0.2249614 ]
 [ 0.73352134  0.32676136]]
[[ 0.90364814  0.73233008]
 [ 0.61350799  0.37857103]
 [ 0.41849363  0.42342687]]
[[ 1.  2.  3.]
 [-3. -7. -1.]
 [ 0.  5. -2.]]
[[ 3.90597057  4.19510365  6.0628376 ]
 [ 4.86557484  6.02541924  3.63600206]]
[[ 0.  0.  0.]
 [ 0.  0.  0.]]
[[ 5.  5.  5.]
 [ 5.  5.  5.]]
[[ 0.04690075  0.34524703  0.48222351]
 [ 0.15054452  0.74442649  0.36477065]]
-38.0
[[-0.5        -0.5        -0.5       ]
 [ 0.15789474  0.05263158  0.21052632]
 [ 0.39473684  0.13157895  0.02631579]]
[[ 1.  0.  0.]
 [ 0.  1.  0.]
 [ 0.  0.  1.]]
(array([-10.65907521,  -0.22750691,   2.88658212]), array([[ 0.21749542,  0.63250104, -0.74339638],
       [ 0.84526515,  0.2587998 ,  0.46749277],
       [-0.4880805 ,  0.73004459,  0.47834331]]))