# hanyuki23/sop

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/hanyuki23-sop).**

5 stars · 0 forks · Python · MIT

## Links

- GitHub: https://github.com/hanyuki23/SoP
- awesome-repositories: https://awesome-repositories.com/repository/hanyuki23-sop.md

## Description

SoP is a universal calibration strategy that resolves multi-target learning conflicts by optimizing each prediction target independently while keeping the backbone frozen. Achieves up to 22% improvement even with simple MLP Plugs.

## Tags

### Part of an Awesome List

- [Forecasting Models](https://awesome-repositories.com/f/awesome-lists/ai/forecasting-models.md) — Non-collective calibrating strategy for forecasting.