In Tensorflow get the names of all the Tensors in a graph

Question

I am creating neural nets with Tensorflow and skflow  for some reason I want to get the values of some inner tensors for a given input  so I am using myClassifier get layer value input   tensorName    myClassifier being a skflow estimators TensorFlowEstimator    However  I find it difficult to find the correct syntax of the tensor name  even knowing its name  and I m getting confused between operation and tensors   so I m using tensorboard to plot the graph and look for the name   Is there a way to enumerate all the tensors in a graph without using tensorboard

User · Answer

There is a way to do it slightly faster than in Yaroslav s answer by using get operations  Here is a quick example   import tensorflow as tf  a   tf constant 1 3  name  const a   b   tf Variable 3 1  name  variable b   c   tf add a  b  name  addition   d   tf multiply c  a  name  multiply    for op in tf get default graph   get operations        print str op name

User · Answer

Since the OP asked for the list of the tensors instead of the list of operations nodes  the code should be slightly different   graph   tf get default graph       tensors per node    node values   for node in graph get operations    tensor names    tensor name for tensors in tensors per node for tensor in tensors

User · Answer

The accepted answer only gives you a list of strings with the names  I prefer a different approach  which gives you  almost  direct access to the tensors   graph   tf get default graph   list of tuples    op values   for op in graph get operations      list of tuples now contains every tensor  each within a tuple  You could also adapt it to get the tensors directly   graph   tf get default graph   list of tuples    op values   0  for op in graph get operations

User · Answer

This worked for me   for n in tf get default graph   as graph def   node      print   n  n

User · Answer

The following solution works for me in TensorFlow 2 3 - def load pb path to pb       with tf io gfile GFile path to pb   rb   as f          graph def   tf compat v1 GraphDef           graph def ParseFromString f read        with tf Graph   as default   as graph          tf import graph def graph def  name             return graph  tf graph   load pb MODEL FILE  sess   tf compat v1 Session graph tf graph     Show tensor names in graph for op in tf graph get operations        print op values     where MODEL FILE is the path to your frozen graph  Taken from here

User · Answer

I ll try to summarize the answers  To get all nodes in the graph   type tensorflow core framework node def pb2 NodeDef   all nodes    n for n in tf get default graph   as graph def   node   To get all ops in the graph   type tensorflow python framework ops Operation   all ops   tf get default graph   get operations    To get all variables in the graph   type tensorflow python ops resource variable ops ResourceVariable   all vars   tf global variables    To get all tensors in the graph   type tensorflow python framework ops Tensor   all tensors    tensor for op in tf get default graph   get operations   for tensor in op values     To get all placeholders in the graph   type tensorflow python framework ops Tensor   all placeholders    placeholder for op in tf get default graph   get operations   if op type   Placeholder  for placeholder in op values     Tensorflow 2 To get the graph in Tensorflow 2  instead of tf get default graph   you need to instantiate a tf function first and access the graph attribute  for example   graph   func get concrete function   graph  where func is a tf function

User · Answer

tf all variables   can get you the information you want   Also  this commit made today in TensorFlow Learn that provides a function get variable names in estimator that you can use to retrieve all variable names easily

User · Answer

Previous answers are good  I d just like to share a utility function I wrote to select Tensors from a graph   def get graph op graph  and conds None  op  and   or conds None          Selects nodes  names in the graph if      - The name contains all items in and conds     - OR AND depending on op     - The name contains any item in or conds      Condition starting with a     are negated      Returns all ops if no optional arguments is given       Args          graph  tf Graph   The graph containing sought tensors         and conds  list str    optional   Defaults to None               and  conditions         op  str  optional   Defaults to  and                How to link the and conds and or conds              with an  and  or an  or          or conds  list str   optional   Defaults to None               or conditions       Returns          list str   list of relevant tensor names             assert op in   and    or        if and conds is None          and conds            if or conds is None          or conds             node names    n name for n in graph as graph def   node       ands             n for n in node names         if all              cond in n if     not in cond             else cond 1   not in n             for cond in and conds                 ors             n for n in node names         if any              cond in n if     not in cond             else cond 1   not in n             for cond in or conds                 if op     and           return               n for n in node names             if n in ands intersection ors                elif op     or           return               n for n in node names             if n in ands union ors              So if you have a graph with ops     model classifier dense kernel    model classifier dense kernel Assign    model classifier dense kernel read    model classifier dense bias    model classifier dense bias Assign    model classifier dense bias read    model classifier dense MatMul    model classifier dense BiasAdd    model classifier ArgMax dimension    model classifier ArgMax     Then running   get graph op tf get default graph      dense     kernel     or     Assign      returns     model classifier dense kernel Assign    model classifier dense bias    model classifier dense bias Assign    model classifier dense bias read    model classifier dense MatMul    model classifier dense BiasAdd

User · Answer

You can do   n name for n in tf get default graph   as graph def   node    Also  if you are prototyping in an IPython notebook  you can show the graph directly in notebook  see show graph function in Alexander s Deep Dream notebook

User · Answer

I think this will do too   print tf contrib graph editor get tensors tf get default graph       But compared with Salvado and Yaroslav s answers  I don t know which one is better

[python] In Tensorflow, get the names of all the Tensors in a graph

Examples related to python

Examples related to tensorflow

Examples related to tensorboard

Examples related to skflow