rapidsai / legate-boost

GBM implementation on Legate
https://rapidsai.github.io/legate-boost/
Apache License 2.0
8 stars 6 forks source link

Implement faster batched prediction #112

Open RAMitchell opened 2 months ago

RAMitchell commented 2 months ago

Implement a static method for the model class, accepting a list of models of that type, that can provide batched prediction for that list of models. In case it is not implemented, just call predict on each model.

A custom batched predict task could be significantly faster than looping over each model, calling predict and summing the predictions.