I am currently using OpenAI version 0.28.1 and trying finetuning davinci. After preparing the training data I used this command for finetuning openai api fine_tunes.create -t training_data.jsonl -m davinci.
This command encountered me the Error: Unknown request URL: POST /v1/fine-tunes. Please check the URL for typos, or see the docs at https://platform.openai.com/docs/api-reference/. (HTTP status code: 404).
From the documentation I learned about new end point but I am wondering how can I fit the endpoint in my command.
You must use a model currently supported, not “davinci” which is a GPT-3 model that is shut off, and only models that use “fine-tuning” are supported ( the prior “fine-tunes” endpoint with a similar name is shut off also).
Confused about the names? The endpoint with “g” is “good” ==> fine-tuning.
Confused about the model to use? Currently, davinci-002 and babbage-002 are available as base models.
You will indeed find under the API reference only the new endpoint, and methods needing the library. If you are on a system that can’t migrate OpenAI Python library, you can make the API calls directly to the endpoint with an http library instead, following the “CURL” demo in the API reference.