CODEWITHSUNDEEP
javascriptaxiosopenai-apifine-tuning
Rémy Noulin
Rémy Noulin

Reputation: 118

OpenAI API error when using a fine-tuned model: "The model `xxxxx` does not exist"

I'm trying to do a chatbot with a fine tuned model.

I'm doing my request like this:

    const API_KEY = "/";
    const ORG_ID = "/";
    const headers = {
      "Content-Type": "application/json",
      Authorization: "Bearer " + API_KEY,
      "OpenAI-Organization": ORG_ID,
    };
    const res = await axios.post(
      "https://api.openai.com/v1/chat/completions",
      {
        model: "ft-modelname",
        messages: [
          {
            role: "system",
            content: "your name is Name.",
          },
          {
            role: "user",
            content: message,
          },
        ],
      },
      { headers }
    );

And after listing my models with the API I can see it, it does exist.

But I have this error in the console:

{
    "error": {
        "message": "The model `ft-modelname` does not exist",
        "type": "invalid_request_error",
        "param": null,
        "code": null
    }
}

After looking up, i can't find a way to do this, some people use open ai engine api instead of completion, but it don't work too.
Do you have any ideas ?

Thank you.

Edit: It works when I'm not using a fine tune model

Upvotes: 0

Views: 1053

Answers (1)

Rok Benko
Rok Benko

Reputation: 22930

UPDATE: 22 August 2023

Fine-tuning for GPT-3.5 is now available, as stated in the official OpenAI blog:

Fine-tuning for GPT-3.5 Turbo is now available, with fine-tuning for GPT-4 coming this fall.

Consequently, the Chat Completions API (i.e., the GPT-3.5 API) endpoint can be used for a fine-tuned model, but only for a GPT-3.5 fine-tuned model. For a GPT-3 fine-tuned model, use the Completions API (i.e., the GPT-3 API) endpoint.

As stated in the official OpenAI documentation:

When a job has succeeded, you will see the fine_tuned_model field populated with the name of the model when you retrieve the job details. You may now specify this model as a parameter to in the Chat Completions API (for gpt-3.5-turbo) or legacy Completions API (for babbage-002 and davinci-002), and make requests to it using the Playground.

Problem

You're using the wrong API endpoint.

Solution

Use the Completions API (i.e., the GPT-3 API) endpoint instead of the Chat Completions API (i.e., the GPT-3.5 API) endpoint.

As stated in the official OpenAI documentation:

When a job has succeeded, the fine_tuned_model field will be populated with the name of the model. You may now specify this model as a parameter to our Completions API, and make requests to it using the Playground.

In general

Python

import os
import openai

openai.api_key = os.getenv("OPENAI_API_KEY")

openai.Completion.create(
    model = FINE_TUNED_MODEL,
    prompt = YOUR_PROMPT)

NodeJS

const { Configuration, OpenAIApi } = require("openai");

const configuration = new Configuration({
  apiKey: process.env.OPENAI_API_KEY,
});

const openai = new OpenAIApi(configuration);

const response = await openai.createCompletion({
  model: FINE_TUNED_MODEL
  prompt: YOUR_PROMPT,
});

cURL

curl https://api.openai.com/v1/completions \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": YOUR_PROMPT, "model": FINE_TUNED_MODEL}'

OpenAI CLI

openai api completions.create -m <FINE_TUNED_MODEL> -p <YOUR_PROMPT>

Your case

Try this:

const res = await axios.post(
  "https://api.openai.com/v1/completions", {
    model: "ft-modelname",
    prompt: "Say this is a test",
  }, {
    headers
  }
);

Upvotes: 1

Related Questions