Tensorflow: How to replace a node in a calculation graph?

by

TL;DR: If you can define the two computations as Python functions, you should do that. If you can’t, there’s more advanced functionality in TensorFlow to serialize and import graphs, which allows you to compose graphs from different sources.

One way to do this in TensorFlow is to build the disjoint computations as separate tf.Graph objects, then convert them to serialized protocol buffers using Graph.as_graph_def():

with tf.Graph().as_default() as g_1:
  input = tf.placeholder(tf.float32, name="input")
  y = f(input)
  # NOTE: using identity to get a known name for the output tensor.
  output = tf.identity(y, name="output")

gdef_1 = g_1.as_graph_def()

with tf.Graph().as_default() as g_2:  # NOTE: g_2 not g_1       
  input = tf.placeholder(tf.float32, name="input")
  z = g(input)
  output = tf.identity(y, name="output")

gdef_2 = g_2.as_graph_def()

Then you could compose gdef_1 and gdef_2 into a third graph, using tf.import_graph_def():

with tf.Graph().as_default() as g_combined:
  x = tf.placeholder(tf.float32, name="")

  # Import gdef_1, which performs f(x).
  # "input:0" and "output:0" are the names of tensors in gdef_1.
  y, = tf.import_graph_def(gdef_1, input_map={"input:0": x},
                           return_elements=["output:0"])

  # Import gdef_2, which performs g(y)
  z, = tf.import_graph_def(gdef_2, input_map={"input:0": y},
                           return_elements=["output:0"]

More Related Contents:

  1. Make a custom loss function in keras
  2. Disable Tensorflow debugging information
  3. Could not load dynamic library ‘cudart64_101.dll’ on tensorflow CPU-only installation
  4. How to run Tensorflow on CPU
  5. How could I use batch normalization in TensorFlow?
  6. NotImplementedError: Layers with arguments in `__init__` must override `get_config`
  7. Keras replacing input layer
  8. Tensorflow: How to write op with gradient in python?
  9. Reproducible results in Tensorflow with tf.set_random_seed
  10. This model has not yet been built error on model.summary()
  11. How can I use a pre-trained neural network with grayscale images?
  12. How to understand the term `tensor` in TensorFlow?
  13. Tensorflow r1.0 : could not a find a version that satisfies the requirement tensorflow
  14. Compute pairwise distance in a batch without replicating tensor in Tensorflow?
  15. TensorFlow ValueError: Cannot feed value of shape (64, 64, 3) for Tensor u’Placeholder:0′, which has shape ‘(?, 64, 64, 3)’
  16. TypeError: ‘Tensor’ object does not support item assignment in TensorFlow
  17. TensorFlow – numpy-like tensor indexing
  18. How do I set TensorFlow RNN state when state_is_tuple=True?
  19. How to know if underfitting or overfitting is occuring?
  20. In Tensorflow, get the names of all the Tensors in a graph
  21. Custom loss function in Keras
  22. Save and load model optimizer state
  23. Keras misinterprets training data shape
  24. In TensorFlow, what is the difference between Session.run() and Tensor.eval()?
  25. Why can I not import Tensorflow.contrib I get an error of No module named ‘tensorflow.python.saved
  26. How do I get the weights of a layer in Keras?
  27. Installing TensorFlow on Windows (Python 3.6.x)
  28. cannot import name ‘_registerMatType’ from ‘cv2.cv2’
  29. TensorFlow: Non-repeatable results
  30. How to do Xavier initialization on TensorFlow

Leave a Comment