openai api fine_tunes.create - not returning the model Id

Soumaly Saha 20

I'm running the below from terminal - openai -k API_KEY api fine_tunes.create -t "train_prepared.jsonl" This is neither throwing any error nor the model Id. It returns "openai 0.28.1" Any suggestion what could be the issue?

AshokPeddakotla-MSFT 30,511 Reputation points

2023-10-09T12:55:52.1633333+00:00

Soumaly Saha Greetings!

Could you confirm if you have also tried specifying model name?
Soumaly Saha 20 Reputation points

2023-10-10T04:39:34.3033333+00:00

Yes , passed this " -m MODEL_NM " as parameter, but no luck.
AshokPeddakotla-MSFT 30,511 Reputation points

2023-11-02T17:24:57.11+00:00

Soumaly Saha Just checking if we are still connected on this discussion? Please let us know if you need further help.

If the response is helpful, please click Accept Answer and Click Yes. So that we can close this thread.

1 answer

AshokPeddakotla-MSFT 30,511 Reputation points

2023-10-12T05:31:54.1166667+00:00

Soumaly Saha This could be due to the fine tuning unavailability for new customers as it is turned off for all regions.

You can see the same on models page as well.

Unfortunately, we don't have any ETA at this moment when fine tuning opens again.

The models page will be updated as soon as we open up regions for fine tuning and users should be able to submit the requests when it is made available.

If you have previously fine-tuned & deployed in a region, then you can fine tune in that region if it's available.

Alternatively, you could use Azure OpenAI on your data feature with Open AI which might remove the need / requirement for fine tuning.

I hope this helps! Let me know if you have any further questions.
Please sign in to rate this answer.
Soumaly Saha 20 Reputation points

2023-10-18T18:04:55.58+00:00

Thanks Ashok for sharing the update. I genuinely feel fine tuning is absolutely needed while training the domain data. Currently, we add information as much as we can in prompts to train domain knowledge & believe , that's the only option we have at this moment to improve the accuracy of the responses. However,this may not be sufficient & quite often produces incorrect responses. I believe,fine tuned model may solve the issue. Look forward to have fine tuning option available in all regions.

AshokPeddakotla-MSFT 30,511 Reputation points

2023-10-19T12:32:24.3366667+00:00

Soumaly Saha Good News. Fine Tuning: is now available with Azure OpenAI Service.

Please see Customize a model with fine-tuning (preview) and Fine Tuning: now available with Azure OpenAI Service for more details.

Do let me know if that helps or have any further queries.
Sign in to comment

Share via

openai api fine_tunes.create - not returning the model Id

1 answer