Can someone help me (with fine-tuning)

I’ve had no success fine-tuning a model since the outage. Right now I’m trying the fine-tuning web interface again but when I load both the training and validation files, I get this error in red:

There was an error uploading the file: Unexpected file format, expected either prompt/completion pairs or chat messages.

I am using the correctly formatted and prepared jsonl file, and why is it saying it expected prompt/completion pairs or chat messages? Thats not how the newest OpenAI documentation said data should be prepared which is using messages/system/user/assistant.

Please, for the love of god, can someone help me. I’ve now spent two days on this.

Just guessing here, since I just started reading the docs, but I remembered this piece: " The conversational chat format is required to fine-tune gpt-3.5-turbo. For babbage-002 and davinci-002, you can follow the prompt completion pair format used for legacy fine-tuning". Maybe it’s the type of model?

I’m fine-tuning gpt3.5-turbo and according to the documentation, the data is supposed to be in this format which is what I’ve done:

{"messages": [{"role": "system", "content": "Marv is a factual chatbot that is also sarcastic."}, {"role": "user", "content": "What's the capital of France?"}, {"role": "assistant", "content": "Paris, as if everyone doesn't know that already."}]}
{"messages": [{"role": "system", "content": "Marv is a factual chatbot that is also sarcastic."}, {"role": "user", "content": "Who wrote 'Romeo and Juliet'?"}, {"role": "assistant", "content": "Oh, just some guy named William Shakespeare. Ever heard of him?"}]}
{"messages": [{"role": "system", "content": "Marv is a factual chatbot that is also sarcastic."}, {"role": "user", "content": "How far is the Moon from Earth?"}, {"role": "assistant", "content": "Around 384,400 kilometers. Give or take a few, like that really matters."}]}

I cannot get any help on this anywhere.

What version of the OpenAI api library are you using?

1 Like

Mattrosine

Have you figured it out yet? If so please tell me and others in the future how to fix this issue.