-
Notifications
You must be signed in to change notification settings - Fork 83
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
Import LightGBM trained model to metarank #1144
Comments
I guess the easiest way would be to copy-paste the most optimal params you've got after the optuna into the lightgbm model settings in metarank's config file and retrain. See this section for details: https://docs.metarank.ai/reference/overview/supported-ranking-models#xgboost-and-lightgbm-backend-options The main issue with importing any trained model is that LGBM has 2.x, 3.x and 4.x formats available and we have to maintain some sort of support for some of these format flavors. It can be a bit tricky if you want to retrain/redeploy your model periodically in an automated way. |
I think could be very valuable, to train a model in an external offline environment, so you can save the model into redis using a new endpoint in your metarank API, so in this API you can control LGBM versions Thanks for your great work!!! |
Thank you for your answer @shuttie. I agree with @ohuarte-matrix. You can put the model store path as a parameter so that users can train, deploy, and update by themself, like this link. With meta rank hyper-params are not enough in my settings:
|
I did many experiments and finetuned with Optuna, then I got the best model of LightGBM.
Could I use that offline trained model and import / integrate it into Metarank?
The text was updated successfully, but these errors were encountered: