C# binding for Menoh DNN inference library.

This project is for 64 bit only.

Nuget

• Menoh

• Prebuild Menoh Binaries(Windows)

• Prebuild MKL-DNN Binaries(Windows)

To execute an example, please put binaries in ./bin/

Visual Studio 2015 or later

mono or Visual Studio For Mac

mono

Double click retrieve_data.bat or execute below command in root directory .

retrieve_data.bat

Open MenohSharp.sln and compile with Visual Studio 2015 or later.

Execute below command in root directory.

sh retrieve_data.sh
xbuild MenohSharp/MenohSharp.csproj /t:build /p:Configuration=Release /p:Platform=x64

Execute ./bin/MenohSharpExample.exe

Execute below command in root directory.

xbuild MenohSharp/MenohSharp.csproj /t:build /p:Configuration=Release /p:Platform=x64
cd menoh
mono MenohSharpExample.exe

If libgdiplus.dylib is not found on Mac, please execute below command (brew and cask are required)

brew cask install mono-mdk

Result is below

-22.45651 -34.58567 -10.29389 24.40432 -0.2879904 -7.913781...
top 5  categories are
8 0.9613832  categories aren01514859 hen
7 0.03693673  categories aren01514668 cock
86 0.001228112  categories aren01807496 partridge
82 0.000224892  categories aren01797886 ruffed grouse, partridge, Bonasa umbellus
97 3.720724E-05  categories aren01847000 drake

Note: retrieve_data.sh downloads data/VGG16.onnx. data/VGG16.onnx is generated by onnx-chainer from pre-trained model which is uploaded at http:https://www.robots.ox.ac.uk/%7Evgg/software/very_deep/caffe/VGG_ILSVRC_16_layers.caffemodel

That pre-trained model is released under Creative Commons Attribution License.

The library license is as follows.

• LICENSE