Cannot use finetuned models in playground

API
1

When trying to use one of my finetuned models in the playground, I am receiving the following error:

Imgur

(you must provide a model parameter)

But I’m sure I’ve selected a model, though:

Imgur

It seems that the playground is currently working with the default models only.

Anyone experiencing something similar?

PS: the finetuned models do work when accessing them through the API

1 Like
2

I’m experiencing the same thing, but accessing the fine-tuned model via an api call does work. Yesterday I did try out the fine tuned model and this worked fine.

3

Same, issue started today

4

I had the same issue, the fine tuned models are all working with the api req and curls, but playground is not recognizing them, must be a bug from the update with chat GPT?

5

Hi from the OpenAI Support team! We’re aware of this and I just submitted a related bug report. Thanks for reporting. I reproduced the issue just now in my own instance of the Playground. Will hope to have an update on a fix soon.

Meanwhile, would you all confirm as a workaround that you’re still able to reliably & successfully complete API calls with your fine tuned models - it’s just the Playground that’s impacted, correct?

2 Likes
6

I can confirm that I’m still able to reliably & successfully complete API calls with my fine tuned models. Thanks for your prompt reaction!

7

Hey @dbk03 and @jeremiah-OpenAI, I had the same issue, but I’m figuring out the API call for the model. I’ve been looking over the docs. Could you paste a template for querying the API?

8

@alexander.mclennan10 I think the best way to figure out the correct API call will still be found in the docs, as it depends a lot on what programming language are you using. Check here:

You will find example calls for multiple programming languages.
Example for python:

import os
import openai
openai.api_key = os.getenv("OPENAI_API_KEY")
openai.Completion.create(
  model="text-davinci-003",
  prompt="Say this is a test",
  max_tokens=7,
  temperature=0
)

But in order to use one of your fine-tuned models, make sure you replace ‘text-davinci-003’ with the ID of one of your fine-tuned models.
You can get a list of your fine-tuned models by sending this call: OpenAI API

@jeremiah-OpenAI Do you have any news on this bug report? The playground still doesn’t work with fine-tuned models. Thanks.

9

I’m experiencing the same problem in the playground.

10

Hi all – apologies for this issue. We just pushed a fix, please refresh the page to give it another try, and let us know if it works!

1 Like
11

I can confirm that the fine-tuned models are now working in Playground, too. Thanks for fixing this!

12

Hello all, I’m still not able to run my model in the playground. Also tried refreshing/reloading the page, facing this issue since last week. Model run when I call an API in Jupyter notebook.

1 Like