Experimental new Python bindings for the VowpalWabbit library
Deploy RL Loop in Azure
Estimators to perform off-policy evaluation
Mature software applications are highly componentized and have well-defined API boundaries among components. This structure leads to constraints making in-process inference challenging due to the difficulty of access to features. The FeatureBroker library presents a solution to this problem. This helped with adoption of VW in inference. We believe that this is a pervasive problem. The goal of open sourcing this library is to reduce the developer friction in adopting VW in software stacks that are built with these constraints.
Vowpal Wabbit is a machine learning system which pushes the frontier of machine learning with techniques such as online, hashing, allreduce, reduct...
最近更新: 7天前Interaction-side integration library for Reinforcement Learning loops: Predict, Log, [Learn,] Update
最近更新: 7天前