1.Tensorflow 数据类型
---------------------------------------------------------------
Data type (dtype) Description
---------------------------------------------------------------
tf.float32 32-bit floating point
tf.float64 64-bit floating point
tf.int8 8-bit signed integer
tf.int16 16-bit signed integer
tf.int32 32-bit signed integer
tf.int64 64-bit signed integer
tf.uint8 8-bit unsigned integer
tf.string String (as bytes array, not Unicode)
tf.bool Boolean
tf.complex64 Complex number, with 32-bit floating point real portion, and 32-bit floating point imaginary portion
tf.qint8 8-bit signed integer (used in quantized Operations)
tf.qint32 32-bit signed integer (used in quantized Operations)
tf.quint8 8-bit unsigned integer (used in quantized Operations)
2.重载符号
TensorFlow重载了一些符号,使得表达更加清晰。
Unary operators
----------------------------------------------------------------------------------------
Operator Related TensorFlow Operation
----------------------------------------------------------------------------------------
-x tf.neg()
~x tf.logical_not()
abs(x) tf.abs()
Binary operators
----------------------------------------------------------------------------------
Operator Related TensorFlow Operation
-----------------------------------------------------------------------------------
x + y tf.add()
x - y tf.sub()
x * y tf.mul()
x / y (Python 2) tf.div()
x / y (Python 3) tf.truediv()
x // y (Python 3) tf.floordiv()
x % y tf.mod()
x ** y tf.pow()
x < y tf.less()
x <= y tf.less_equal()
x > y tf.greater()
x >= y tf.greater_equal()
x & y tf.logical_and()
x | y tf.logical_or()
x ^ y tf.logical_xor()
3.图和会话
# 一般我们写程序的时候,都会指定是哪个图
# 得到默认图
default_graph = tf.get_default_graph()
g = tf.Graph()
with g.as_default():
a = tf.constant(2)
b = tf.constant(4)
c = tf.add(a,b)
with default_graph.as_default():
aa = tf.constant(4)
bb = tf.constant(8)
cc = tf.add(aa, bb)
sess = tf.Session(graph = g)
# 这个语句是对的,因为会话里面的图是g
print(sess.run(c))
# 这个语句是错误的,因为只有图g能执行,而图default_graph不行,因为不在会话里
print(sess.run(cc))