Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
HY-SOAR:Self-Correction for Optimal Alignment and Refinement in Diffusion Models
| Date | Stars |
|---|---|
| 2026-07-31 | 741 |
| 2026-08-05 | 741 |
| 2026-08-06 | 741 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center">
# HY-SOAR: Self-Correction for Optimal Alignment and Refinement in Diffusion Models
</div>
<div align="center">
<a href=https://arxiv.org/abs/2604.12617 target="_blank"><img src=https://img.shields.io/badge/Report-b5212f.svg?logo=arxiv height=22px></a>
<a href=https://github.com/Tencent-Hunyuan/HY-SOAR target="_blank"><img src=https://img.shields.io/badge/Code-2ecc71?logo=github&logoColor=black height=22px></a>
<a href=https://hy-soar.github.io/ target="_blank"><img src=https://img.shields.io/badge/HY.SOAR-1abc9c?logo=homeassistant&logoColor=white height=22px></a>
<a href=https://x.com/TencentHunyuan target="_blank"><img src=https://img.shields.io/badge/Hunyuan-black.svg?logo=x height=22px></a>
</div>
<p align="center">
<img src="assets/soar1.png" width="60%">
</p>
**Beyond SFT and RL: Self-Correction during Generation without Reward Models, Preference Labels, or Negative Samples.**
## 🔥 News
- **April 2026**: 🎉 HY-SOAR open source - Training and evaluation code publicly available.
## 🗂️ Contents
- [🔥 News](#-news)
- [📖 Introduction](#-introduction)
- [✨ Key Features](#-key-features)
- [🖼 Showcases](#-showcases)
- [📑 Open-Source Plan](#-open-source-plan)
- [🛠 Environment Setup](#-environment-setup)
- [🎯 Reward Preparation](#-reward-preparation)
- [🚀 Usage](#-usage)
- [📊 Evaluation](#-evaluation)
- [🧾 Data Format](#-data-format)
- [📚 Citation](#-citation)
- [🙏 Acknowledgement](#-acknowledgement)
---
## 📖 Introduction
**HY-SOAR** (Self-Correction for Optimal Alignment and Refinement) is a reward-free post-training method for rectified-flow diffusion models. It targets exposure bias in the denoising trajectory: standard SFT trains the denoiser on ideal forward-noising states from real data, while inference conditions on states produced by the model's own earlier predictions. Once an early denoising step drifts, later steps must recover from states that were not directly optimized, so errors can compound across the trajectory.
Instead of waiting for a terminal reward after a full rollout, SOAR teaches the model to correct its own trajectory errors at the timestep where they occur. Given a clean latent $z_0$, noise endpoint $z_1$, and condition $c$, SOAR:
1. Samples an on-trajectory noisy state and performs one stop-gradient CFG rollout step with the current model
2. Re-noises the resulting off-trajectory state toward the same noise endpoint $z_1$ to create auxiliary states
3. Supervises the denoiser with the analytical correction target $v_{\mathrm{corr}} = (z_{\sigma_{t'}} - z_0) / \sigma_{t'}$
This gives SOAR an on-policy, dense, and reward-free training signal. The base objective subsumes standard SFT, while the auxiliary correction loss trains on nearby model-induced states, making SOAR a stronger first post-training stage that remains compatible with subsequent reward-based alignment.
<p align="center">
<img src="assets/soar.png" alt="HY-SOAR Teaser"/>
</p>
## ✨ Key Features
* 🧭 **Exposure-Bias Correction:** SOAR directly addresses the mismatch between ground-truth training states and model-induced inference states, the source of many compounding denoising failures.
* 🔁 **On-Policy Off-Trajectory Supervision:** Off-trajectory states are produced by the current model's own rollout, so the training distribution co-evolves with the model instead of staying fixed to the SFT data trajectory.
* 🎯 **Reward-Free Dense Objective:** SOAR requires no reward model, preference labels, or negative samples. It provides per-timestep correction supervision and avoids terminal-reward credit assignment.
* 📐 **Geometric Correction Target:** Re-noising uses the same noise endpoint as the base flow-matching pair, keeping auxiliary states near the original transport ray and yielding a concrete correction velocity anchored to $z_0$.
* 🔧 **Compatible Post-Training Stage:** The SOAR loss extends the standard flow-matching objective, so it can replace SFT as a stronger first post-Excerpt of 13,856 characters
Read on GitHub
14
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:0b7db6adad0e118d, desc:alignment