Hi
I was trying to make a chatbot by openai api for myself, because chatGPT is so amazing.
So I was trying to Fine-Tune at first and then CompletionText, but I found some problems:
1.After Fine-Tune, a new model will be generated every time, but I have to use the new model to perform CompletionText, otherwise CompletionText can not use my newest Fine-Tune training data.
2.If I use new model to perform CompletionText, CompletionText can not use my old Fine-Tune training data, is there a way to just update Fine-Tune training data for one customize model?
3. Does ChatGPT use CompletionText to show result as well? Why ChatGPT answers questions can be a chatbot has its own logical, but If I use CompletionText api, its answer just like completing my questions and the Fine-Tune completion?