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

feat: Ability to set additional inference parameters such as top_p and num_ctx #36

Open
sammcj opened this issue Nov 14, 2024 · 0 comments · May be fixed by #37
Open

feat: Ability to set additional inference parameters such as top_p and num_ctx #36

sammcj opened this issue Nov 14, 2024 · 0 comments · May be fixed by #37

Comments

@sammcj
Copy link

sammcj commented Nov 14, 2024

It would be great if you could set common inference parameters such as top_p and num_ctx.

  • top_p helps control quality vs creativity (much smarter than temperature)
  • min_p another, newer algorithm for quality vs creativity (some would argue even smarter than top_p)
  • num_ctx configures what context size to load the model with
  • repetition_penalty configures how much repetition is allowed in the generations (this is an oversimplification but it's useful)

I think they're the most commonly used parameters, however a complete list is available https://github.com/ollama/ollama/blob/main/docs/modelfile.md#valid-parameters-and-values

@sammcj sammcj linked a pull request Nov 14, 2024 that will close this issue
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant