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

server: allow setting penalize-nl on server webpage #6431

Merged
merged 1 commit into from
Apr 4, 2024
Merged

server: allow setting penalize-nl on server webpage #6431

merged 1 commit into from
Apr 4, 2024

Conversation

sha224
Copy link
Contributor

@sha224 sha224 commented Apr 2, 2024

This change allows users to set penalize-nl on the server webpage which fixes #6430 .

The user can set penalize-nl by using a checkbox shown in the following screenshot.
Screenshot_20240401_231105

@github-actions GitHub Actions

This comment was marked as off-topic.

Sign in to view
phymbert
phymbert requested changes Apr 2, 2024
View reviewed changes
Copy link
Collaborator

@phymbert phymbert left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please use spaces instead of tabs. See the linter issues

@sha224 sha224 force-pushed the server-penalize-nl branch 2 times, most recently from 9044c20 to a6c07e3 Compare April 2, 2024 10:07
@sha224
Copy link
Contributor Author

sha224 commented Apr 2, 2024

I have addressed the tab issue.

@Jeximo
Copy link
Contributor

Jeximo commented Apr 2, 2024 •
edited
Loading

For certain models, the chat causes infinite newlines after a few turns.

I'm a bit confused because all #6127 did was set repeat-pentalty to 1.0(disabled) whereas server docs show penalize-nl defaults to true... Is the readme wrong?

Does setting repeat-penalty to default 1.1 fix #6430?

Either way, I like the addition of a penalize-nl button within server.

Edit: Ah, further checking and the server readme is outdated because #6127 also set penalize-nl to false, and there's no command line to change the value to true.

@sha224 sha224 force-pushed the server-penalize-nl branch 2 times, most recently from 8ab82cd to 8cb2a59 Compare April 3, 2024 00:32
@sha224 sha224 force-pushed the server-penalize-nl branch from 8cb2a59 to 6f7ab29 Compare April 3, 2024 01:02
@sha224 sha224 requested a review from phymbert April 3, 2024 01:15
ngxson
ngxson approved these changes Apr 4, 2024
View reviewed changes
Copy link
Collaborator

@ngxson ngxson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. I'll wait for @phymbert to review once more and merge it.

@phymbert phymbert merged commit 2e66913 into ggerganov:master Apr 4, 2024
58 checks passed
@phymbert
Copy link
Collaborator

phymbert commented Apr 4, 2024

@ngxson thanks, please note the new step in Server CI Verify server deps dedicated for this kind of PR.

tybalex pushed a commit to rubra-ai/tools.cpp that referenced this pull request Apr 17, 2024
@sha224 @tybalex
server: allow penalizing repetition of newlines on server webpage (gg…
9e6a3d5 
…erganov#6431)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@ngxson ngxson ngxson approved these changes

@phymbert phymbert phymbert approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Server: Infinite newlines during chat on server webpage
4 participants
@sha224 @Jeximo @phymbert @ngxson