Step 1: compile MegEngine for deployment following README.md
Step 2: compile the example by
$CXX -o xor_deploy -I$MGE_INSTALL_PATH/include xor_deploy.cpp -L$MGE_INSTALL_PATH/lib64/ -lmegengine
where $CXX
is the C++ compiler and $MGE_INSTALL_PATH
is the MegEngine install path.
Step 3: run with dumped model
The dumped model can be obtained by running xornet.py
LD_LIBRARY_PATH=$MGE_INSTALL_PATH/lib64:$LD_LIBRARY_PATH ./xor_deploy xornet_deploy.mge 0.6 0.9
Sample output:
Predicted: 0.999988 1.2095e-05