-
Notifications
You must be signed in to change notification settings - Fork 3.2k
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
Make MaxTokens optional #1362
Comments
anthonypuppo
changed the title
Make MaxTokens an optional argument
Make MaxTokens optional
Jun 7, 2023
alexchaomander
added
.NET
Issue or Pull requests regarding .NET code
kernel
Issues or pull requests impacting the core kernel
labels
Jun 8, 2023
Thanks for making the PR! And for the great discussion about this during Office Hours! |
github-merge-queue bot
pushed a commit
that referenced
this issue
Jul 8, 2023
### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Some LLM API's have the "max tokens" parameter as optional (OpenAI, Azure OpenAI, etc.) and will default to the remaining tokens left in the models context window. This PR changes the default value of max tokens from an explicitly defined amount to null. Note that skill specific max token amounts are unchanged, this PR aims to just change the defaults for library consumers right now. Fixes #1362 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Change max tokens to a nullable int type (`int` to `int?`). ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [✔️] The code builds clean without any errors or warnings - [✔️] The PR follows SK Contribution Guidelines (https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) - [✔️] The code follows the .NET coding conventions (https://learn.microsoft.com/dotnet/csharp/fundamentals/coding-style/coding-conventions) verified with `dotnet format` - [✔️] All unit tests pass, and I have added new tests where possible - [✔️] I didn't break anyone 😄 Co-authored-by: Lee Miller <lemiller@microsoft.com> Co-authored-by: Shawn Callegari <36091529+shawncal@users.noreply.github.com> Co-authored-by: Roger Barreto <19890735+RogerBarreto@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Some LLM's, such as ChatGPT, default to using the remaining context window if the maximum amount of tokens is not included in the request. For such models it would be beneficial to not pass the max tokens argument every time if the intention is to get the most elaborate response possible. This is of course possible by calculating remaining tokens, but is a pain (and performance hit) to constantly calculate the remaining tokens (assuming you even get it right).
The text was updated successfully, but these errors were encountered: