ARM64 SnapDragon X Elite Dell Laptop, VS2022, C# - Headaches and deadends

[StewartWoody]

Hi,

I am struggling to implement some OpenAI functions using 2.2.0-beta.4 with instantiation thus:

using OpenAI;
using OpenAI.Chat;

...

public OpenAIClient Ai { get; private set; }

and some basic Hello World...

 public async Task<string> ChatAsync(string userText,
                                    CancellationToken ct = default)
{
    var req = new ChatCompletionRequest
    {
        Model    = _model,
        Messages =
        {
            ChatMessage.FromSystem(_systemPrompt),
            ChatMessage.FromUser(userText)
        }
    };

    var resp = await _client.ChatEndpoint.GetCompletionAsync(req, ct);
    return resp.Choices[0].Message.Content.Trim();
}

}

Getting a raft of unknown commands around the ChatCompletionRequest, ChatMessage, and ChatEndpoint.

Wondering if there's a root cause around my ARM64 CPU ?

Managed to get a half baked solution using legacy code but also now falling over with Pinecone (vector database) and LangChain .NET for the RAG.

Any experiences around this with some working code perhaps?

[joseharriaga]

Thank you for reaching out, @StewartWoody ! Could you share what other packages you are using? The code snippet that you included here doesn't look like our library. ChatCompletionRequest, ChatEndpoint, etc. are not classes in the OpenAI package.

The examples here might help:
🔗 https://github.com/openai/openai-dotnet/tree/main/examples/Chat

[StewartWoody]

Thankyou so much for your prompt reply. That answer speaks volumes. ChatGPT now claims "ChatCompletionRequest, ChatMessage, or ChatEndpoint—those examples came from a future pre-release that you don’t have." Sounds like it is developing an ego ; ) I will review the demo code and work with that instead.

Cheers!