lain¶
Models for osu! that learn from user replays to predict scores.
This project exists to build better recommendation systems for osu! as well as explore how different users play osu!.
Lain is currently being used to power: combine
Error Model¶
The goal of the error model is to learn the user’s tendencies by looking at maps on the detail individual hit objects. The error model tries to predict the user’s expected accuracy by learning what their expected aim and accuracy error will be on any given object.
For more information see the docs.
Contents: