Is there any multi-round conversation dataset maker?

Community
,
1

When we build a multi-round conversation dataset, we nede to paste the Q&A prompts, which is time-consuming. Is there any tools for better making it?

2 Likes
2

what’s the definition of that?

3

A dataset for fine-tuning, like:

{"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."}]}
2 Likes
4

Got it … thanks …could you use Excel (or Libreoffice) I wonder, and just enter the conversation and have it build the correct data format?

(it would make sense for that tool to already exist :slight_smile: )

1 Like
5

I have same idea as you… But there isn’t… Maybe everyone thinks in that way

2 Likes
6

There are various labelling tools, not sure how suitable they are for this particular use case, but I know that Label Studio has support for creating Q&A datasets (disclaimer: I haven’t used it myself for this though).

1 Like
7

i use datasetai.com

can test finetunes right there

1 Like