tensorgraph

Tensorgraph is an TensorFlow example to show

• How to generate checkpoint, graph.pb, tensorboard.
• How to use checkpoint and graph.pb to merge to freeze graph. (not finished yet)
• How to load graph with tensorflow c++ api and do the prediction.

More detailed description in my blog post

Requirement

• tensorflow installation, https://www.tensorflow.org/
  go to "GET STARTED" --> "installing from source"
• bazel installation, http://www.bazel.io/docs/install.html
  

gengraph

How to generate checkpoint, graph.pb, tensorboard.
The directory struct is

mnist.py
board/
models/

After run

$ python mnist.py

The directory struct will be expected to

mnist.py
board/
    event.out.tfevents
models/
    graph.pb
    model.ckpt
Mnist_data/
    ...

generate frozen graph

From Tensorflow official guide says that:

What this does is load the GraphDef, pull in the values for all the variables from the latest checkpoint file, and then replace each Variable op with a Const that has the numerical data for the weights stored in its attributes It then strips away all the extraneous nodes that aren't used for forward inference, and saves out the resulting GraphDef into an output file

Hence, we do the following steps to generate frozen graph

bazel build tensorflow/python/tools:freeze_graph && \
bazel-bin/tensorflow/python/tools/freeze_graph \
--input_graph=graph.pb \
--input_checkpoint=model.ckpt \
--output_graph=/tmp/frozen_graph.pb --output_node_names=softmax

loadgraph

How to load graph with tensorflow c++ api and do the prediction.
Put the directory to tensorflow source code. Here is the final directory structure:

tensorflow/tensorflow/loadgraph
tensorflow/tensorflow/loadgraph/mnist.cc
tensorflow/tensorflow/loadgraph/MNIST.h
tensorflow/tensorflow/loadgraph/BUILD

Compile and Run

From inside the TensorFlow project folder call $bazel build tensorflow/tensorflow/loadgraph:mnistpredict From the repository root, go into bazel-bin/tensorflow/loadgraph. Copy the frozen_graph.pb and Mnist_data to bazel-bin/tensorflow/loadgraph Then run ./mnistpredict and check the output

Reference

MNIST_Loader
Load graph with tensorflow c++ api