Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BUG]Output my input #692

Open
Yummy416 opened this issue Jan 15, 2024 · 4 comments
Open

[BUG]Output my input #692

Yummy416 opened this issue Jan 15, 2024 · 4 comments
Labels
pending Something isn't working

Comments

@Yummy416
Copy link

After I finished fine-tuning with the full parameters, instead of using your chatbot, I called it in the way of pipeline. Why would I output my input in answer every time? For example:
The input is: Please tell me who you are.
The output is: Please tell me who you are. I'm a big AI model.

@Yummy416 Yummy416 added the pending Something isn't working label Jan 15, 2024
@research4pan
Copy link
Contributor

Thanks for your interest in LMFlow! The behavior of the model depends on the model type, training parameters and adopted dataset. Could you please provide the following information so we could better understand the issue?

  • model name
  • an example of data sample
  • prompt structure
  • end string/character

Thanks very much 😄

@Yummy416
Copy link
Author

Yummy416 commented Jan 17, 2024

@research4pan Hello, here are the details:
model name: Baichuan2-13B-Chat
example of data sample:
image
end string/character: "</s>"
I trained in this way, input content, he will input and output all the content to me output

@research4pan
Copy link
Contributor

According to my knowledge, Baichuan is a decoder-only model. So it is better to use "text-only" format as follows

###Human: 请给我写一个... ###Answer: 你能帮我...

and use chatbot with --prompt_structure ###Human: {input_text} ###Answer: and --end_string ###.

Hope that can answer your question 😄

@Yummy416
Copy link
Author

Ok, thanks for your reply. I'll have a try

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
pending Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants