TensorFlowTOP

A highly experimental TensorFlow plugin for TouchDesigner.

This will be updated.

Setup

1. Clone this repository.
2. Create a folder inside called libs.
3. Clone the TensorFlow repository inside of libs.
4. Follow the instructions below to build TensorFlow.

Building TensorFlow from Source

1. Install Visual Studio 2015.
2. Install CUDA 8.0 and cuDNN 5.1.
3. Install Anaconda for Windows.
4. Run conda install python=3.5 to install the proper version of Python (you can verify that the correct version of Python is installed by opening a new cmd prompt and running the command python: you should see text like Python 3.5.4 ... before the interpreter is launched).
5. Install SWIG (latest .exe for Windows x64).
6. Add the SWIG folder location to your PATH, as it is referenced during the build.
7. Navigative to libs\tensorflow\tensorflow\contrib\cmake\.
8. Create a folder here called build.
9. Navigative inside this folder.
10. Run the command: cmake .. -A x64 -DCMAKE_BUILD_TYPE=Release: this will generate a Visual Studio 2015 solution file called tensorflow.sln.
11. Open this file in Visual Studio 2015.
12. By default, the project named ALL_BUILD will be selected.
13. Make sure that your configuration is set to Release and x64.
14. Right-click ALL_BUILD in the solution explorer and click build.
15. Wait about 2 hours...
16. Finally, if the build succeeds, create an environment variable $TENSORFLOW_BUILD pointing to the build directory.

Creating an Example C++ Project

1. Create a new Visual Studio 2015 project.
2. Add a new file named main.cpp.
3. Under C/C++ -> General -> Additional Include Directories, add:

$(TENSORFLOW_BUILD);
$(TENSORFLOW_BUILD)\..\..\..\..\;
$(TENSORFLOW_BUILD)\..\..\..\..\third_party\eigen3;
$(TENSORFLOW_BUILD)\external\zlib_archive;
$(TENSORFLOW_BUILD)\external\gif_archive\giflib-5.1.4;
$(TENSORFLOW_BUILD)\external\png_archive;
$(TENSORFLOW_BUILD)\external\jpeg_archive;
$(TENSORFLOW_BUILD)\external\lmdb;
$(TENSORFLOW_BUILD)\external\eigen_archive;
$(TENSORFLOW_BUILD)\gemmlowp\src\gemmlowp;
$(TENSORFLOW_BUILD)\jsoncpp\src\jsoncpp;
$(TENSORFLOW_BUILD)\external\farmhash_archive;
$(TENSORFLOW_BUILD)\external\farmhash_archive\util;
$(TENSORFLOW_BUILD)\external\highwayhash;
$(TENSORFLOW_BUILD)\cub\src\cub;
$(TENSORFLOW_BUILD)\external\nsync\public;
$(TENSORFLOW_BUILD)\protobuf\src\protobuf\src;
$(TENSORFLOW_BUILD)\re2\install\include;
$(TENSORFLOW_BUILD)\external\sqlite;
$(TENSORFLOW_BUILD)\grpc\src\grpc\include;
$(TENSORFLOW_BUILD)\snappy\src\snappy;

3. Under C/C++ -> Preprocessor -> Preprocessor Definitions, add:

COMPILER_MSVC;
PLATFORM_WINDOWS;

4. Under Linker -> General -> Additional Library Directories, add $(TENSORFLOW_BUILD).
5. Under Linker -> Input -> Additional Dependencies, add the following libraries:

re2\src\re2\$(Configuration)\re2.lib;
grpc\src\grpc\Release\grpc++_unsecure.lib;
grpc\src\grpc\Release\grpc_unsecure.lib;
grpc\src\grpc\Release\gpr.lib;
zlib\install\lib\zlibstatic.lib;
gif\install\lib\giflib.lib;
png\install\lib\libpng12_static.lib;
jpeg\install\lib\libjpeg.lib;
lmdb\install\lib\lmdb.lib;
jsoncpp\src\jsoncpp\src\lib_json\$(Configuration)\jsoncpp.lib;
farmhash\install\lib\farmhash.lib;
fft2d\src\lib\fft2d.lib;
highwayhash\install\lib\highwayhash.lib;
nsync\install\lib\nsync.lib;
snappy\src\snappy\Release\snappy.lib;
protobuf\src\protobuf\Release\libprotobuf.lib;
tf_cc.dir\Release\tf_cc.lib;
tf_cc_ops.dir\Release\tf_cc_ops.lib;
tf_cc_framework.dir\Release\tf_cc_framework.lib;
tf_core_cpu.dir\Release\tf_core_cpu.lib;
tf_core_direct_session.dir\Release\tf_core_direct_session.lib;
tf_core_framework.dir\Release\tf_core_framework.lib;
tf_core_kernels.dir\Release\tf_core_kernels.lib;
tf_core_lib.dir\Release\tf_core_lib.lib;
tf_core_ops.dir\Release\tf_core_ops.lib;
tf_cc_while_loop.dir\Release\tf_cc_while_loop.lib;
Release\tf_protos_cc.lib;
sqlite\install\lib\sqlite.lib

which are all relative to the $(TENSORFLOW_BUILD) path. 6. Under Linker -> Command Line -> Additional Options, add:

/machine:x64 
/ignore:4049 
/ignore:4197 
/ignore:4217 
/ignore:4221 
/WHOLEARCHIVE:tf_cc.lib
/WHOLEARCHIVE:tf_cc_framework.lib
/WHOLEARCHIVE:tf_cc_ops.lib
/WHOLEARCHIVE:tf_core_cpu.lib
/WHOLEARCHIVE:tf_core_direct_session.lib
/WHOLEARCHIVE:tf_core_framework.lib
/WHOLEARCHIVE:tf_core_kernels.lib
/WHOLEARCHIVE:tf_core_lib.lib
/WHOLEARCHIVE:tf_core_ops.lib 
/WHOLEARCHIVE:libjpeg.lib

to the text field 7. Back in main.cpp, add the following include directives:

#include "tensorflow/cc/ops/const_op.h"
#include "tensorflow/cc/ops/image_ops.h"
#include "tensorflow/cc/ops/standard_ops.h"
#include "tensorflow/core/framework/graph.pb.h"
#include "tensorflow/core/framework/tensor.h"
#include "tensorflow/core/graph/default_device.h"
#include "tensorflow/core/graph/graph_def_builder.h"
#include "tensorflow/core/lib/core/errors.h"
#include "tensorflow/core/lib/core/stringpiece.h"
#include "tensorflow/core/lib/core/threadpool.h"
#include "tensorflow/core/lib/io/path.h"
#include "tensorflow/core/lib/strings/stringprintf.h"
#include "tensorflow/core/platform/init_main.h"
#include "tensorflow/core/platform/logging.h"
#include "tensorflow/core/platform/types.h"
#include "tensorflow/core/public/session.h"
#include "tensorflow/core/util/command_line_flags.h"

8. Generate the model with export_model.py.

References

• https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/cmake/README.md
• https://joe-antognini.github.io/machine-learning/build-windows-tf
• http://anonimousindonesian.blogspot.com/2017/12/tutorial-how-to-build-tensorflow.html
• https://github.com/jhjin/tensorflow-cpp
• https://gist.github.com/kyrs/9adf86366e9e4f04addb