Wills Zhuang
Wills Zhuang

Reputation: 11

Deploying Tensorflow Trained model on HoloLens

I am developing an extension for an existing app for HoloLens that should use a model which was trained by tensorflow.

My attempt is to use TensorFlowSharp because the existing app is using unity (C#).

I build using:

Scripting Runtime Version: .Net 4x or Equivalent
Scripting Backend: IL2CPP
API Compatibility Level: .Net Standard 2.0
TensorflowSharp 1.7
Tensorflow 1.7.1
unity2018
Windows 10
libtensorflow.dll (build from windows 32bit, as hololens is in 32 bit)

My Question is what is this error and any idea how to solve it?

It seems like something happen in the IL2CPP backend but no extra information is given.

I have succeeded in building the apps with unity, however I get the error when I deploy it to hololens:

Exception thrown at 0x77452D72 in TFSharpTest2018.exe: Microsoft C++ 
exception: Il2CppExceptionWrapper at memory location 0x0E5FF36C.

Upvotes: 1

Views: 403

Answers (0)

Related Questions