Build your application

You must link your application with the Arm NN SDK library, the TensorFlow Lite parsing library, the inference test library that is part of ArmNN SDK, and the Boost library.

The following command compiles and links your application with the necessary libraries:

$(CXX) -DARMNN_TF_LITE_PARSER \
-I$(ARMNN_ROOT)/include \
-I$(ARMNN_ROOT)/src/backends \
-I$(ARMNN_ROOT)/src/armnnUtils \
-I$(ARMNN_ROOT)/tests \
-I$(BOOST_ROOT)/include \
-Wall -O3 -std=c++14 -fPIE mobilenetv1_quant_tflite.cpp -o mobilenetv1_quant_tflite -pie \
-L$(ARMNN_BUILD)-L$(ARMNN_BUILD)/tests \
-L$(BOOST_ROOT)/lib \
-larmnn -larmnnTfLiteParser -linferenceTest \
-lboost_system -lboost_filesystem -lboost_program_options

The commands are specified in the Makefile of the example code. Alternatively, you can build the example code with the make command. Refer to the README document from the example code for more information.

A successful build of your application produces the mobilenetv1_quant_tflite executable file. 

Note

The example code runs on the Android platform. Therefore if you are on an X86 host, $(CXX) must specify the same cross-platform compiler that you use to build ArmNN SDK.

Previous Next