using bitnetcpp with c# throught a wrapper to compiled bitnetcpp

[a-trax]

Hi.
I'm looking for a way to call bitnetcpp in a C# console application.
I successfully compiled bitnetcpp, and i have llama.dll, ggml.dll, llava_shared.dll and a lot of llama-*.exe

Is it possible to make a c++/CLI wrapper around it to call it from a C# console application ?

For now, i can load the dll with my wrapper, but i can't find a method to call to make inferences.

the code of my wrapper :

#include "pch.h"
#include "BitNetWrapper.h"
#define WIN32_LEAN_AND_MEAN
#include <Windows.h>
#include <msclr/marshal_cppstd.h>
#include <string>

using namespace msclr::interop;
using namespace System;

namespace BitNetWrapper {

    typedef const char* (__cdecl* LLAMA_INFER_FUNC)(const char*);

    HMODULE hLlamaDll = nullptr;
    LLAMA_INFER_FUNC llama_infer = nullptr;

    InferenceEngine::InferenceEngine(String^ dllPath) {
        std::string nativePath = marshal_as<std::string>(dllPath);
        hLlamaDll = LoadLibraryA(nativePath.c_str());

        if (!hLlamaDll) {
            throw gcnew Exception("Can't load llama.dll");
        }

        llama_infer = (LLAMA_INFER_FUNC)GetProcAddress(hLlamaDll, "llama_infer");

        if (!llama_infer) {
            throw gcnew Exception("Function 'llama_infer' not found in llama.dll");
        }
    }

    String^ InferenceEngine::Infer(String^ prompt) {
        std::string nativePrompt = marshal_as<std::string>(prompt);
        const char* result = llama_infer(nativePrompt.c_str());
        return gcnew String(result);
    }

    InferenceEngine::~InferenceEngine() {
        if (hLlamaDll) {
            FreeLibrary(hLlamaDll);
            hLlamaDll = nullptr;
        }
    }
}

And the code of the console application using this wrapper :

            var engine = new InferenceEngine("llama.dll");
            string result = engine.Infer("What is the capital of England ?");
            Console.WriteLine(result);
            Console.Read();

This code load llama.dll and his dependencies, but it crashes with "Function 'llama_infer' not found in llama.dll"

Thanks