-
Notifications
You must be signed in to change notification settings - Fork 1.2k
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
Feature: Configure RPM for specific models #764
Comments
Thanks for trying aider and filing this issue. Aider should have retried that error a bunch of times before finally giving up? Aider doesn't use the litellm proxy, just the python library. And I don't know what the proxy would do if the client exceeds the rate limit? Probably just return a rate limit error just like google is? |
Unfortunately, it stucks with the error
Or sometimes with the error like
and I need to press Ctrl+C two times and re-run aider. Perhaps, moving to litellm or just adding retry mechanism after certain point of time would help it.
It should just wait for some time to make the request again based on the RPM from the specific model quota and just using exponential backoff.
It obligates to re-run aider manually again and again and it's not suitable. Better to have auto-retry based on the allowed RPM for the model or just based on the exponential backoff. |
Aider does retry litellm.RateLimitError. If all the retries fail, only then does it report the error to the user. |
I'm going to close this issue for now, but feel free to add a comment here and I will re-open or file a new issue any time. |
Issue
Hi! First of all, thank you for such unique tool.
I wonder, is it possible to set request per minute amount? For example, if I use free version of Gemini, it allows 2 RPM https://ai.google.dev/pricing . So, I am getting the error now.
Since aider uses litellm, it will be greate to pass litellm settings from some YAML file and use rpm attribute, for example, like this https://litellm.vercel.app/docs/proxy/reliability#step-1---set-deployments-on-config
Please, make an attention to this, since it will help to bring more delicate fine tuning.
Thank you!.
Version and model info
Aider v0.40.6
Model: gemini/gemini-1.5-pro-latest with diff-fenced edit format
Git repo: .git with 6 files
Repo-map: using 1024 tokens
The text was updated successfully, but these errors were encountered: