update

.gitattributes

@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 assets/method_overview.png filter=lfs diff=lfs merge=lfs -text
 assets/qualitative_gallery.jpg filter=lfs diff=lfs merge=lfs -text
 assets/training_curves.png filter=lfs diff=lfs merge=lfs -text
+assets/opsd_v2.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -24,8 +24,9 @@ tags:
 [![Paper](https://img.shields.io/badge/arXiv-2608.24646-b31b1b?logo=arxiv)](https://arxiv.org/abs/2608.24646)
 [![Project Page](https://img.shields.io/badge/Project-Page-3B82F6)](https://diffusionopsd.github.io/)
 [![Code](https://img.shields.io/badge/Code-GitHub-181717?logo=github)](https://github.com/worldbench/DiffusionOPSD)
+[![V2 Blog](https://img.shields.io/badge/V2-Blog-B0448C?logo=rss&logoColor=white)](https://diffusionopsd.github.io/blog/v2/)
 
-<img src="assets/qualitative_gallery.jpg" width="100%" alt="Images generated with DiffusionOPSD">
+<a href="https://diffusionopsd.github.io/blog/v2/"><img src="assets/opsd_v2.png" width="100%" alt="DiffusionOPSD v2 training overview"></a>
 
 </div>
 
@@ -35,13 +36,39 @@ tags:
 
 By turning image-level rewards into explicit, continually refreshed intermediate supervision, DiffusionOPSD makes **target construction** and **finite realization** separately observable. Across SD3.5-M and Z-Image-Turbo, it achieves the best final held-out score in **19 of 20** reward-matched settings and reduces training GPU-hours relative to DiffusionNFT by **40%** and **63%**, respectively.
 
+<p align="center">
+  <img src="assets/qualitative_gallery.jpg" width="100%" alt="Images generated with DiffusionOPSD">
+</p>
+
 <p align="center">
   <img src="assets/method_overview.png" width="100%" alt="DiffusionOPSD method overview">
 </p>
 
 ## Released Checkpoints
 
-This repository provides nine rank-32 LoRA adapters:
+### DiffusionOPSD\* (v2): `new.zip`
+
+[`new.zip`](./new.zip) contains the checkpoints for **DiffusionOPSD\***, the accelerated v2 recipe. It keeps the same target-construction principle and uses one reward-ascent step, broader prompt coverage, and rollouts that stop at the query. See the [v2 blog](https://diffusionopsd.github.io/blog/v2/) for the update.
+
+The archive contains **eight rank-32 LoRA adapters**, covering **SD3.5-M and Z-Image-Turbo × HPSv3, HPSv2.1, CLIPScore, and PickScore**. All are the final **`checkpoint-100` EMA inference weights**. The archive is **1.46 GB**.
+
+```bash
+hf download WeiChow/DiffusionOPSD new.zip --local-dir checkpoints/diffusionopsd-star
+unzip checkpoints/diffusionopsd-star/new.zip -d checkpoints/diffusionopsd-star
+```
+
+The extracted `opsd_v2_2x4_20261012/` directory contains:
+
+| Backbone | Adapter folders |
+|---|---|
+| SD3.5-M | `sd35-m-hpsv3`, `sd35-m-hpsv2`, `sd35-m-clipscore`, `sd35-m-pickscore` |
+| Z-Image-Turbo | `z-image-turbo-hpsv3`, `z-image-turbo-hpsv2`, `z-image-turbo-clipscore`, `z-image-turbo-pickscore` |
+
+Each folder contains `adapter_config.json` and `adapter_model.safetensors`.
+
+### Original DiffusionOPSD adapters
+
+The nine top-level adapter folders below contain the original DiffusionOPSD checkpoints. The **DiffusionOPSD\*** checkpoints are packaged separately in `new.zip`.
 
 | Checkpoint | Backbone | Training objective |
 |---|---|---|
@@ -55,7 +82,7 @@ This repository provides nine rank-32 LoRA adapters:
 | [`z-image-turbo-hpsv3`](./z-image-turbo-hpsv3) | [Z-Image-Turbo](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo) | HPSv3 |
 | [`z-image-turbo-pointwise`](./z-image-turbo-pointwise) | [Z-Image-Turbo](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo) | Pointwise reward |
 
-Download all released adapters with:
+Download the repository, including `new.zip`, with:
 
 ```bash
 hf download WeiChow/DiffusionOPSD --local-dir checkpoints/diffusionopsd
@@ -70,6 +97,7 @@ hf download WeiChow/DiffusionOPSD --local-dir checkpoints/diffusionopsd
 - **Paper:** [On-Policy Self-Distillation in Diffusion Models](https://arxiv.org/abs/2608.24646)
 - **Code:** [worldbench/DiffusionOPSD](https://github.com/worldbench/DiffusionOPSD)
 - **Project page:** [diffusionopsd.github.io](https://diffusionopsd.github.io/)
+- **V2 blog:** [DiffusionOPSD v2](https://diffusionopsd.github.io/blog/v2/)
 
 Please refer to the [GitHub repository](https://github.com/worldbench/DiffusionOPSD) for installation, inference, evaluation, and training instructions.
 

new.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89b166aa38877f55aa5462e828f07e6c42d9588671ef63a5b22e446cc201a1a0
+size 1457897214