python - OpenAI GPT-3 API : Which file formats can be used for fine-tuning?

Question

随着我们进入 AI 的动荡时代。我也把我的水洒进了海里。因为我是 pythonian，所以所有的尝试都是在 python/anaconda 中完成的。

是否有人对可传递给 GPT AI 家族的“数据格式”有一些经验？

在文档中推荐使用 OpenAI 工具进行控制。后跟文档推荐格式(“提示:”，“完成:”)字符串标记为:

  ["str" = in quotes,"/" = separator ,"@>" = unique symbol, 
   " " = comp. starts with empty space]

  'Prompt':    'Hello AI..!!/@>' 
  'Completion': ' How are you today?/@>' 

“完成”应该在每个字符串的开头都有空白。到目前为止，我只能找到简单的例子:

Col1             Col2
'Prompt':        'Completion':
'Text/@>'        ' Text/@>'

它有什么办法可以理解更复杂的数据集吗？效果更暗淡。数据框？示例:

     Col1        Col2             Col3         Col4        
    'Prompt_a':  'Completion_a':  'Prompt_b':  'Completion_b':
    'Text/@>'    ' Text/@>'       'Text/@>'    ' Text/@>

较长的上下文文本是作为“str/@>”传递的，还是需要一些分区？

' text text text /@>'

非常感谢所有的回答和提前的努力。

已检查:https://help.openai.com/en/articles/6811186-how-do-i-format-my-fine-tuning-data

Answer 1

如官方所述OpenAI documentation :

Your data must be a JSONL document, where each line is aprompt-completion pair corresponding to a training example. You canuse our CLI data preparation tool to easily convert your data intothis file format.

This tool accepts different formats, with the only requirement thatthey contain a prompt and a completion column/key. You can pass a CSV,TSV, XLSX, JSON or JSONL file, and it will save the output into aJSONL file ready for fine-tuning, after guiding you through theprocess of suggested changes.