When I call openai API to access the generate davinci or chatgpt model, it works as expected.
However, when I call the fine-tuned davinci version below error returned, although the model has not changed:
The server is currently overloaded with other requests. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if the error persists.
The same behaviour happen from the playgroud.
We are also having issues with receiving response from the fine tuned models with davinci
The issue is resolved; I wonder why the tuned models get lower priority access than the default models.
I don’t think it’s that they have less priority, but they’re likely on their own server. If I remember correctly, if you don’t use the fine-tune model frequently, it can sometimes take a bit to “spin-up,” and become more responsive. I’m sure a lot of people have tons of fine-tuning experiment models, so keeping them all live all the time isn’t really practical, so I’m betting that if one isn’t used, it’s shuffled down to a lower priority because of its low usage?
Then again, it could just be normal server problems. Going from 1 mil to 100 mil users in a couple months is not easy, I’m sure.
Glad you got it sorted. Thanks for coming back to let us know.