This is a runtime library for inference Neural Network created by Neural Network Libraries.
Neural Network Libraries and Neural Network Console can output DNN training result as NNP file.
File Format Converter can convert NNP file into NNB file interpreted by Runtime or C Source Code to call Functions.
This library has the following features.
- It is almost independent from external libraries and is written in Pure C (C99).
- almost mean it depends C standard math library.
- It has been developed with priority over readability rather than performance, making it ideal for learning and porting.
- It adopts an extensible architecture, and you can use the function you implemented yourself as necessary for applications that need performance.
- Provides CLI utility that can infer NNB file and dump NNB to confirm network I/O or function list.
For Linux
The following is required as prerequisite: cmake git C compiler
For example, on Ubuntu you can install these dependencies with the following command.
$ sudo apt install build-essential cmake git
Then, you can build nnabla-c-runtime with following command.
$ git clone https://github.com/sony/nnabla-c-runtime.git
$ cd nnabla-c-runtime
$ mkdir build
$ cd build
$ cmake ..
$ cmake --build .
After compilation is successfully finished, you can exececute nnablart.exe as following:
$ ./src/nnablart/nnablart version
C Runtime Version [1.2.0.dev1_c1]
NNB Version [3]
NNB Revision [API_LEVEL_6]
For Windows
The following is required as prerequisite:
cmake installed and in your PATH git installed and in your PATH Visual C++ compiler available.
The easiest way to install is to use Chocolatey. Once you set up Chocolatey,
> choco install cmake git choco visualstudio2019-workload-vctools
It can be installed with the above command only. Then,
Open Visual C++ 2019 x64 Native Build Tools Command Prompt from windows menu.
> git clone https://github.com/sony/nnabla-c-runtime.git
> cd nnabla-c-runtime
> mkdir build
> cd build
> cmake ..
> cmake --build . --config Release
After compilation is finished successfully, You can exececute nnablart.exe
>src\nnablart\Release\nnablart.exe version
C Runtime Version [1.2.0.dev1_c1] NNB Version [3] NNB Revision [API_LEVEL_6]
./nnablart infer bin_class.nnb input.bin output_1 in Linux
./nnablart.exe infer bin_class.nnb input.bin output_1 in Win