Tensor定义

tensor张量可以理解为n维数组:

• 0维张量是一个数(Scalar/number),
• 1维张量是向量(Vector),
• 2维张量是矩阵(Martrix),
• 以此类推…

基础运算

import tensorflow as tf
a=tf.add(3,5)
print(a)

Tensor("Add:0", shape=(), dtype=int32)

TF的加法方法，但是通常的赋值并不是正常运行加法。
需要将被赋值的变量a放入session运行才能看到运算结果。

a=tf.add(3,5)
sess=tf.Session()
print(sess.run(a))
sess.close()

8

将运算结果存入sess稍后再用的写法

a=tf.add(3,5)
with tf.Session() as sess:
    print(sess.run(a))

8

tf.Session()封装了一个执行运算的环境，用tensor对象内的赋值进行运算

混合运算

x=2
y=3
op1 =tf.add(x,y)
op2=tf.multiply(x,y)
op3=tf.pow(op2,op1)
with tf.Session() as sess:
    op3=sess.run(op3)
    print(op3)

7776

Subgraphs

x=2
y=3
add_op=tf.add(x,y)
mul_op=tf.multiply(x,y)
useless=tf.multiply(x,add_op)
pow_op=tf.pow(add_op,mul_op)
with tf.Session() as sess:
    z=sess.run(pow_op)
    print(z)

15625

由于求Z值并不需要计算useless部分，所以session并没有计算它

x=2
y=3
add_op=tf.add(x,y)
mul_op=tf.multiply(x,y)
useless=tf.multiply(x,add_op)
pow_op=tf.pow(add_op,mul_op)
with tf.Session() as sess:
    z,not_useless=sess.run([pow_op,useless])
    print(z)
    print(not_useless)

15625
10

同时进行两个计算

Graph

g=tf.Graph()
with g.as_default():
    x=tf.add(3,5)
    
sess=tf.Session(graph=g)
with tf.Session() as sess: #此处两行的打包方式已经过时，如果报错需要改成下面的格式
    sess.run(g)

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in __init__(self, fetches, contraction_fn)
    270         self._unique_fetches.append(ops.get_default_graph().as_graph_element(
--> 271             fetch, allow_tensor=True, allow_operation=True))
    272       except TypeError as e:


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\ops.py in as_graph_element(self, obj, allow_tensor, allow_operation)
   3034     with self._lock:
-> 3035       return self._as_graph_element_locked(obj, allow_tensor, allow_operation)
   3036 


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\ops.py in _as_graph_element_locked(self, obj, allow_tensor, allow_operation)
   3123       raise TypeError("Can not convert a %s into a %s." % (type(obj).__name__,
-> 3124                                                            types_str))
   3125 


TypeError: Can not convert a Graph into a Tensor or Operation.


During handling of the above exception, another exception occurred:


TypeError                                 Traceback (most recent call last)

<ipython-input-20-5c5906e5d961> in <module>()
      5 sess=tf.Session(graph=g)
      6 with tf.Session() as sess:
----> 7     sess.run(g)


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in run(self, fetches, feed_dict, options, run_metadata)
    887     try:
    888       result = self._run(None, fetches, feed_dict, options_ptr,
--> 889                          run_metadata_ptr)
    890       if run_metadata:
    891         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _run(self, handle, fetches, feed_dict, options, run_metadata)
   1103     # Create a fetch handler to take care of the structure of fetches.
   1104     fetch_handler = _FetchHandler(
-> 1105         self._graph, fetches, feed_dict_tensor, feed_handles=feed_handles)
   1106 
   1107     # Run request and get response.


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in __init__(self, graph, fetches, feeds, feed_handles)
    412     """
    413     with graph.as_default():
--> 414       self._fetch_mapper = _FetchMapper.for_fetch(fetches)
    415     self._fetches = []
    416     self._targets = []


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in for_fetch(fetch)
    240         if isinstance(fetch, tensor_type):
    241           fetches, contraction_fn = fetch_fn(fetch)
--> 242           return _ElementFetchMapper(fetches, contraction_fn)
    243     # Did not find anything.
    244     raise TypeError('Fetch argument %r has invalid type %r' %


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in __init__(self, fetches, contraction_fn)
    273         raise TypeError('Fetch argument %r has invalid type %r, '
    274                         'must be a string or Tensor. (%s)'
--> 275                         % (fetch, type(fetch), str(e)))
    276       except ValueError as e:
    277         raise ValueError('Fetch argument %r cannot be interpreted as a '


TypeError: Fetch argument <tensorflow.python.framework.ops.Graph object at 0x0000018E70371A20> has invalid type <class 'tensorflow.python.framework.ops.Graph'>, must be a string or Tensor. (Can not convert a Graph into a Tensor or Operation.)

教程的例子有报错，需要改成下面的格式

g=tf.Graph()
with g.as_default():
    x=tf.add(3,5)
    

with tf.Session(graph=g) as sess: #将上面的两行改成一行
    sess.run(x)                   #不能直接运行graph

g=tf.Graph()
with g.as_default():
a=3
b=5
x=tf.add(a,b)
sess = tf.Session(graph=g)
sess.close()

向graph内添加加法运算，并且设为默认graph

g1=tf.get_default_graph()
g2=tf.graph()
#将运算加入到默认graph
with g1.as_default():
    a=tf.Constant(3)       #不会报错，但推荐添加到自己创建的graph里
    
#将运算加入到用户创建的graph
with g2.as_default():
    b=tf.Constant(5)

建议不要修改默认graph

** Graph 的优点 **

• 节省运算资源，只计算需要的部分
• 将计算分解为更小的部分
• 让分布式运算更方便，向多个CPU，GPU或其它设备分配任务
• 适合那些使用directed graph的机器学习算法

Graph 与 Session 的区别

• Graph定义运算，但不计算任何东西，不保存任何数值，只存储你在各个节点定义的运算。
• Session可运行Graph或一部分Graph，它负责在一台或多台机器上分配资源，保存实际数值，中间结果和变量。

下面通过以下例子具体阐明二者的区别：

graph=tf.Graph()
with graph.as_default():#每次TF都会生产默认graph,所以前两行其实并不需要
    variable=tf.Variable(42,name='foo')
    initialize=tf.global_variables_initializer()
    assign=variable.assign(13)

创建变量，初始化值42，之后赋值13

graph=tf.Graph()
with graph.as_default():#每次TF都会生产默认graph,所以前两行其实并不需要
    variable=tf.Variable(42,name='foo')
    initialize=tf.global_variables_initializer()
    assign=variable.assign(13)

with tf.Session(graph=graph) as sess:  
    sess.run(initialize)      #记得将计算步骤在此处列出来
    sess.run(assign)
    print(sess.run(variable))

13

定义的计算数量达到三个时就要使用graph。但是variable每次运算都要在session内run一遍，如果跳过此步骤，就无法获取运算后变量数值。（也就相当于没计算过）

graph=tf.Graph()
with graph.as_default():#每次TF都会生产默认graph,所以前两行其实并不需要
    variable=tf.Variable(42,name='foo')
    initialize=tf.global_variables_initializer()
    assign=variable.assign(13)

with tf.Session(graph=graph) as sess:
    print(sess.run(variable))    #未列出计算步骤所以报错

---------------------------------------------------------------------------

FailedPreconditionError                   Traceback (most recent call last)

~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _do_call(self, fn, *args)
   1322     try:
-> 1323       return fn(*args)
   1324     except errors.OpError as e:


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _run_fn(session, feed_dict, fetch_list, target_list, options, run_metadata)
   1301                                    feed_dict, fetch_list, target_list,
-> 1302                                    status, run_metadata)
   1303 


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\errors_impl.py in __exit__(self, type_arg, value_arg, traceback_arg)
    472             compat.as_text(c_api.TF_Message(self.status.status)),
--> 473             c_api.TF_GetCode(self.status.status))
    474     # Delete the underlying status object from memory otherwise it stays alive


FailedPreconditionError: Attempting to use uninitialized value foo
     [[Node: _retval_foo_0_0 = _Retval[T=DT_INT32, index=0, _device="/job:localhost/replica:0/task:0/device:CPU:0"](foo)]]


During handling of the above exception, another exception occurred:


FailedPreconditionError                   Traceback (most recent call last)

<ipython-input-25-cb7c04ce65af> in <module>()
      6 
      7 with tf.Session(graph=graph) as sess:
----> 8     print(sess.run(variable))


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in run(self, fetches, feed_dict, options, run_metadata)
    887     try:
    888       result = self._run(None, fetches, feed_dict, options_ptr,
--> 889                          run_metadata_ptr)
    890       if run_metadata:
    891         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _run(self, handle, fetches, feed_dict, options, run_metadata)
   1118     if final_fetches or final_targets or (handle and feed_dict_tensor):
   1119       results = self._do_run(handle, final_targets, final_fetches,
-> 1120                              feed_dict_tensor, options, run_metadata)
   1121     else:
   1122       results = []


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _do_run(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)
   1315     if handle is None:
   1316       return self._do_call(_run_fn, self._session, feeds, fetches, targets,
-> 1317                            options, run_metadata)
   1318     else:
   1319       return self._do_call(_prun_fn, self._session, handle, feeds, fetches)


~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\client\session.py in _do_call(self, fn, *args)
   1334         except KeyError:
   1335           pass
-> 1336       raise type(e)(node_def, op, message)
   1337 
   1338   def _extend_graph(self):


FailedPreconditionError: Attempting to use uninitialized value foo
     [[Node: _retval_foo_0_0 = _Retval[T=DT_INT32, index=0, _device="/job:localhost/replica:0/task:0/device:CPU:0"](foo)]]

graph=tf.Graph()
with graph.as_default():#每次TF都会生产默认graph,所以前两行其实并不需要
    variable=tf.Variable(42,name='foo')
    initialize=tf.global_variables_initializer()
    assign=variable.assign(13)

with tf.Session(graph=graph) as sess:  
    sess.run(initialize)      #计算步骤，列到第几步就计算到第几步
    print(sess.run(variable))

42