Unable to run LLamaSharp provider on CPU

Describe the bug

I'm trying to run the example code without GPU.
In the console I see the following error:

llama_model_load_from_file_impl: invalid value for main_gpu: 0 (available devices: 0)

Using LLamaSharp directly I can get around this error by setting the GpuLayerCount property of ModelParams to 0.

Steps to reproduce the bug

I've tried the following code:

// get model path
var modelPath = HuggingFaceModelDownloader.GetModelAsync(
    repository: "TheBloke/Thespis-13B-v0.5-GGUF",
    fileName: "thespis-13b-v0.5.Q2_K.gguf",
    version: "main");
    
modelPath.Wait();

    // load model
var model = new LLamaSharpModelInstruction(new LLamaSharpConfiguration
        {
//            MainGpu = -1,
            PathToModelFile = modelPath.Result,
            Temperature = 0,
            GpuLayerCount = 0
        }).UseConsoleForDebug();

// building a chain
var prompt = @"
You are an AI assistant that greets the world.
World: Hello, Assistant!
Assistant:";

var chain =
    Set(prompt, outputKey: "prompt")
    | LLM(model, inputKey: "prompt");

chain.RunAsync().Wait();

Expected behavior

The code should also run without a GPU.

Screenshots

No response

NuGet package version

    <PackageReference Include="LLamaSharp" Version="0.23.0" />
    <PackageReference Include="LLamaSharp.Backend.Cpu" Version="0.23.0" />
    <PackageReference Include="LLamaSharp.semantic-kernel" Version="0.23.0" />
    <PackageReference Include="LLamaSharp.kernel-memory" Version="0.23.0" />
    <PackageReference Include="LangChain" Version="0.17.0" />
    <PackageReference Include="LangChain.Providers.LLamaSharp" Version="0.17.0" />

Additional context

No response

Start with the reproduced code using LLamaSharpModelInstruction and LLamaSharpConfiguration, then run it with the LLamaSharp CPU backend and GpuLayerCount set to 0. Compare the provider's configuration with direct LLamaSharp behavior; done means the supplied example runs without a GPU and no invalid main_gpu error occurs.