Faster Tree Models

We offer 10x faster prediction with XGBoost or LightGBM compared to native libraries or Treelite on CPUs, with NO changes to your model, so you can deploy a 10x bigger and more accurate model for the same latency.

Check out our blog post and open-source benchmarking repo for more information!

 

Contact us about your use case! (Response within 24 hrs)

Thanks for submitting!