GPT-3 Fine-tuning: How to generate "JSONL" training file without code?

Need help
1

Hello! I need help from you no-coders.
As the title says, I try to fine-tunig GPT-3 for OpenAI, but I’m having trouble generating the
JSONL training data without code in the first step “Prepare training data”.
I used this converter to convert CSV to JSONL, but when I request the file endpoint in Postman, it returns the following error.

  • part of my training data
{"prompt":"My stomach is growling","completion":"Would you like something to eat?"}
{"prompt":"I feel tired","completion":"That is a concern. If there is anything I can do for you, please let me know."}
  • Error message
{
    "error": {
        "message": "Expected file to have JSONL format, where every line is a JSON dictionary. Line 1 is not a dictionary (HINT: line starts with: \"{\"p...\").",
        "type": "invalid_request_error",
        "param": null,
        "code": null
    }
}

It says that there are some newlines at the beginning of the file. But the content is fine according to the JONLint.

What should I do? I am eagerly awaiting your advice.
Thank you.

1 Like
2

Did you try running the API call and postman?

3

Thank you @AliFarahat .

Yes, I received the above error message in Postman.

4

Have you tried wrapping each line in quotes as suggested by the message in the error?

5

Thank you, @dorilama .

Yes, I tried with data each line wrapped in ["] or ["backslash], but the same error message was returned.

6

So it’s probably the way you make the request with the API connector.

7

【Solved.】
I changed Unicode from UTF-8 with Bom to UTF-8 without BOM, so the file upload was successful.
It seemed to be related to the fact that I was creating training data in Japanese…
Thanks everyone.

1 Like
8

I’m having this issue as well, but I’m not sure how to change Unicode.

9

Any help on this would be great, as I’m still getting that error even after converting to jsonl and trying to upload

10

Hi @jamesgetsresults . How about then?

You can convert unicode with a text editor for developers, like Adobe Dreamweaver.

The last item, " Unicode Signature (BOM)", in this page is relevant. Please translate and use accordingly!

11

But Dreamweaver isn’t automated - it would have to be manual - yea?

12

You are right. I have not found a way to automate.

13

This topic was automatically closed after 70 days. New replies are no longer allowed.