Yijun Liu commited on
Commit
3eb7599
·
1 Parent(s): ac26578

Add resource badges and standardize the paper citation

Browse files
Files changed (2) hide show
  1. README.md +20 -5
  2. citation.bib +8 -0
README.md CHANGED
@@ -3,20 +3,30 @@ license: apache-2.0
3
  base_model:
4
  - Qwen/Qwen3-VL-8B-Instruct
5
  pipeline_tag: image-text-to-text
 
 
6
  tags:
7
  - reward-model
8
  - text-to-image
9
  - human-preference
10
  - rlhf
 
11
  ---
12
 
13
  # HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities
14
 
 
 
 
 
 
 
 
15
  HPSv3++ is a **capability-aware and RL-iteration-aware** text-to-image (T2I) reward model, built on the `Qwen/Qwen3-VL-8B-Instruct` backbone with a Capability Encoder, a FiLM conditioning head, and a three-layer RankNet reward head.
16
 
17
  A Capability Encoder implicitly infers the generative ability of the model that produced an image, while the RL iteration step is supplied as an explicit condition; the two are jointly modulated through FiLM so that a single reward model produces calibrated preference scores across generators of differing capability and different stages of RL optimization.
18
 
19
- The training/evaluation dataset, HPDv3++, is released separately: [Junjun2333/HPDv3-PlusPlus](https://huggingface.co/datasets/Junjun2333/HPDv3-PlusPlus).
20
 
21
  ## Files
22
 
@@ -35,10 +45,15 @@ The training/evaluation dataset, HPDv3++, is released separately: [Junjun2333/HP
35
 
36
  ## Citation
37
 
 
 
38
  ```bibtex
39
- @misc{hpsv3pp,
40
- title = {HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities},
41
- author = {HPSv3++ Team},
42
- year = {2026}
 
 
 
43
  }
44
  ```
 
3
  base_model:
4
  - Qwen/Qwen3-VL-8B-Instruct
5
  pipeline_tag: image-text-to-text
6
+ datasets:
7
+ - Junjun2333/HPDv3-PlusPlus
8
  tags:
9
  - reward-model
10
  - text-to-image
11
  - human-preference
12
  - rlhf
13
+ - arxiv:2606.14657
14
  ---
15
 
16
  # HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities
17
 
18
+ <p align="center">
19
+ <a href="https://arxiv.org/abs/2606.14657"><img src="https://img.shields.io/badge/arXiv-2606.14657-b31b1b.svg" alt="Paper: arXiv 2606.14657"></a>
20
+ <a href="https://github.com/PlantPotatoOnMoon/HPSv3-PlusPlus"><img src="https://img.shields.io/badge/GitHub-Code-181717?logo=github" alt="GitHub: HPSv3++ code"></a>
21
+ <a href="https://huggingface.co/datasets/Junjun2333/HPDv3-PlusPlus"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Dataset-HPDv3%2B%2B-yellow" alt="Dataset: HPDv3++"></a>
22
+ <a href="https://huggingface.co/Junjun2333/HPSv3-PlusPlus/resolve/main/citation.bib?download=true"><img src="https://img.shields.io/badge/Citation-BibTeX-2563eb" alt="Citation / BibTeX"></a>
23
+ </p>
24
+
25
  HPSv3++ is a **capability-aware and RL-iteration-aware** text-to-image (T2I) reward model, built on the `Qwen/Qwen3-VL-8B-Instruct` backbone with a Capability Encoder, a FiLM conditioning head, and a three-layer RankNet reward head.
26
 
27
  A Capability Encoder implicitly infers the generative ability of the model that produced an image, while the RL iteration step is supplied as an explicit condition; the two are jointly modulated through FiLM so that a single reward model produces calibrated preference scores across generators of differing capability and different stages of RL optimization.
28
 
29
+ **[HPDv3++](https://huggingface.co/datasets/Junjun2333/HPDv3-PlusPlus)** is an independently usable human preference training dataset and benchmark with approximately **212K preference pairs** covering text fidelity and aesthetic quality. See the [dataset guide](https://github.com/PlantPotatoOnMoon/HPSv3-PlusPlus#dataset-hpdv3).
30
 
31
  ## Files
32
 
 
45
 
46
  ## Citation
47
 
48
+ Please cite the paper when using the HPSv3++ model or HPDv3++ dataset. [Download BibTeX](https://huggingface.co/Junjun2333/HPSv3-PlusPlus/resolve/main/citation.bib?download=true).
49
+
50
  ```bibtex
51
+ @article{liu2026hpsv3plusplus,
52
+ title = {HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities},
53
+ author = {Liu, Yijun and Huang, Jie and Xue, Zeyue and Li, Yuming and He, Ruizhe and Li, Haoran and Ge, Shijia and Fu, Siming},
54
+ journal = {arXiv preprint arXiv:2606.14657},
55
+ year = {2026},
56
+ doi = {10.48550/arXiv.2606.14657},
57
+ url = {https://arxiv.org/abs/2606.14657}
58
  }
59
  ```
citation.bib ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ @article{liu2026hpsv3plusplus,
2
+ title = {HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities},
3
+ author = {Liu, Yijun and Huang, Jie and Xue, Zeyue and Li, Yuming and He, Ruizhe and Li, Haoran and Ge, Shijia and Fu, Siming},
4
+ journal = {arXiv preprint arXiv:2606.14657},
5
+ year = {2026},
6
+ doi = {10.48550/arXiv.2606.14657},
7
+ url = {https://arxiv.org/abs/2606.14657}
8
+ }