Support for Chat Completions

[rsthn]

Today chat completions are now available:
https://platform.openai.com/docs/api-reference/chat/create

Uses a new end-point (https://api.openai.com/v1/chat/completions), and the latest model "gpt-3.5-turbo".

[megalon]

I am taking a stab at this, but I've hit a wall.
The API is throwing an error that the endpoint is incorrect, but it looks right to me.

System.Net.Http.HttpRequestException: Error at chat/completions (https://api.openai.com/v1/chat/completions) with HTTP status code: NotFound. Content: {
  "error": {
    "message": "Invalid URL (POST /v1/chat/completions)",
    "type": "invalid_request_error",
    "param": null,
    "code": null
  }
}

Anybody have any suggestions for why this might be happening?
I am not very familiar with how APIs work to be honest.

Here are my changes so far
master...megalon:OpenAI-API-dotnet:feature/chat

Edit;
I was able to get this to work. The error was in the data in the request.

[davidyack]

@megalon confirming your changes worked for me

[rsthn]

This is awesome! Thank you @megalon

[dexinhuang]

Can someone provide an example of calling the new ChatGPT completion?
I used to use this code, var c = new CompletionRequest(memory.ToString(), GetModel(model), MaxTokens, Temperature, null, Top_p, NumOutputs, PresencePenalty, FrequencyPenalty, LogProbs,null, StopSequences); var results = await OAIEngine.Instance.Api.Completions.CreateCompletionsAsync(c); And I'm not sure how to achieve the same results in the new format

[rsthn]

For testing I'm currently using:

var req = new OpenAI_API.Chat.ChatRequest();
req.Model = OpenAI_API.Models.Model.ChatGPTTurbo;
req.Messages = new System.Collections.Generic.List<OpenAI_API.Chat.ChatMessage>();
req.Messages.Add(new OpenAI_API.Chat.ChatMessage("user", "Give me a list of 10 colors"));
req.Temperature = 0;
req.MaxTokens = 100;

var result = await client.Chat.CreateChatAsync(req);
foreach (var i in result.Choices)
{
    Console.WriteLine("["+i.FinishReason+"] " + i.Message.Role + " => " + i.Message.Content);
}

[dexinhuang]

Amazing, thanks!

[dexinhuang]

Quick question, I'm being charged for gpt-3.5-turbo-0301, but I've even removed the model from the model list. Is anyone else running into the same issue?

[megalon]

gpt-3.5-turbo-0301 is the only "chat" model available from the OpenAi API right now. gpt-3.5-turbo uses the latest model, and since 0301 is the only model available, that's the one it uses. "Developers who use the gpt-3.5-turbo model will always get our recommended stable model, while still having the flexibility to opt for a specific model version. For example, today we’re releasing gpt-3.5-turbo-0301, which will be supported through at least June 1st, and we’ll update gpt-3.5-turbo to a new stable release in April." https://openai.com/blog/introducing-chatgpt-and-whisper-apis If you want to use a different model like davinci, you need to use the old "completions" API endpoint

…

On Sat, Mar 4, 2023, 7:31 AM dexinhuang ***@***.***> wrote: Quick question, I'm being charged for gpt-3.5-turbo-0301, but I've even removed the model from the model list. Is anyone else running into the same issue? — Reply to this email directly, view it on GitHub <#54 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGTOITONJ4EEHG3W2KRSVWLW2NN3NANCNFSM6AAAAAAVMQVFGU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[dexinhuang]

Great! I appreciate the information! Thanks!

[ericliu0408]

Could you advise that when it could come out?