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.

Sign up for GitHub

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

Jump to bottom

Calibration Dataset: how to avoid computing loss on instructions? #525

Open
RanchiZhao opened this issue Jun 28, 2024 · 0 comments
Open

Calibration Dataset: how to avoid computing loss on instructions? #525

RanchiZhao opened this issue Jun 28, 2024 · 0 comments

Comments

@RanchiZhao
Copy link

I would like to know about chat models. When I use AWQ for calibration, I do not want to compute the loss for the instructions, but only for the responses. I want to know how to handle this when inputting the calibration dataset. For example, how should I handle the labels and attention masks?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@RanchiZhao