You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Absolutely love treelite. Was able to fit an entire XGBoost model into a 5MB package for AWS Lambda.
But...there is no "tl2cgen-runtime" the way you had a "treelite-runtime". I had to make my own by "stubbing out" scipy and sklearn. Otherwise it imports literally hundreds of megs of libraries and so cannot fit into an AWS Lambda package.
It seems like it would only be a few hours of work to either lazy-load or make a "tl2cgen-runtime" so that I can use the model without needing all the heavyweight stuff that generates the model.
Thanks for all the work on this, it's an amazing accomplishment...
The text was updated successfully, but these errors were encountered:
Absolutely love treelite. Was able to fit an entire XGBoost model into a 5MB package for AWS Lambda.
But...there is no "tl2cgen-runtime" the way you had a "treelite-runtime". I had to make my own by "stubbing out" scipy and sklearn. Otherwise it imports literally hundreds of megs of libraries and so cannot fit into an AWS Lambda package.
It seems like it would only be a few hours of work to either lazy-load or make a "tl2cgen-runtime" so that I can use the model without needing all the heavyweight stuff that generates the model.
Thanks for all the work on this, it's an amazing accomplishment...
The text was updated successfully, but these errors were encountered: