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

Reducing overhead of Eco-CI #62

Open

ArneTR opened this issue Apr 12, 2024 · 0 comments

Labels

enhancement good first issue

Member

ArneTR commented Apr 12, 2024

At the moment the XGBoost model is re-learned every time a measurement step is executed.

In order to have it quicker we can two two steps:

Learn the model only one time per Github Action Runner instantiation and then pickle the current memory structure so we can execute it later directly. Inferencing cost is in the low ms range.
Or we could even pre-learn the model for all the different CPU architectures on Github, as there are only a few. THis way we only have to run through model re-training everytime we do a new release

Help wanted :)

ArneTR added enhancement good first issue labels

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment