How to use Tensorflow trained Machine Learning Model in iOS

Question

I am trying to recognise image which has been trained using Tensorflow. I followed this steps and finally i got success to train my own dataset and its give a good prediction result. All of the code is in python. Now i am trying to use this trained model in my iOS project . I am following this tutorial to use my trained model in my iOS project. But when i followed those steps i got an error in my mac terminal window like -

"ERROR: /tensorflow/tensorflow/core/kernels/BUILD:2235:1: C++ compilation of rule '//tensorflow/core/kernels:self_adjoint_eig_v2_op' failed: gcc failed: error executing command /usr/bin/gcc -U_FORTIFY_SOURCE -fstack-protector -Wall -B/usr/bin -B/usr/bin -Wunused-but-set-parameter -Wno-free-nonheap-object -fno-omit-frame-pointer -g0 -O2 '-D_FORTIFY_SOURCE=1' -DNDEBUG ... (remaining 124 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 4. gcc: internal compiler error: Killed (program cc1plus)" .

What will be the solution for this issue or is there any way to convert this tensor-flow model to iOS supported Core ML Model ? Here i'm sharing a screen shot of that error. Please help me out. Thanks.

Answer 1

Most likely this is due to insufficient memory inside the docker container for bazel to run with default parameters. Try rerunning (after bazel clean ) the build with these extra flags to the build command: --local_resources 2048,2.0,1.0 -j 1 . You can also try giving a little more resources: --local_resources 4096,2.0,1.0 -j 1

Here are some links for this issue: https://github.com/tensorflow/tensorflow/issues/1530 Tensorflow Serving Compile Error Using Docker on OSX