Qwen2.5-VL-7B-GRPO-Composition-Score-Class

zhiyuanyou/Qwen2.5-VL-7B-GRPO-Composition-Score-Class

Qwen2.5-VL-7B-GRPO-Composition-Score-Class is a image-text-to-text model from zhiyuanyou. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. The card lists the license as mit.

Composition Assessment Model ( project page / codes / paper ) weights fully fine-tuned on AVA, CADB, GAIC, and KUPCP datasets.

Downloads · 30 days

2.5K

66% of all-time downloads

All-time downloads

3.8K

Public

Parameters

8.3B

16.6 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors16.6 GB · 100%

At a glance

Task: Image-Text-to-Text
License: mit
Model type: qwen2_5_vl
Access: Public
Created: Apr 17, 2026
Updated: Apr 17, 2026
SHA: 75d7da9f

Try a prompt

Base models

Qwen/Qwen2.5-VL-7B-Instruct

Model card

From the Hugging Face model README

PhotoFramer Composition Assessment Model

Composition Assessment Model ( project page / codes / paper ) weights fully fine-tuned on AVA, CADB, GAIC, and KUPCP datasets.

Composition Assessment (PLCC / SRCC) and Classification (Accuracy) Results

The metrics for composition assessment is SRCC / PLCC, and the metric for composition classification is accuracy.

Dataset	CADB Assessment	GAIC Assessment	AVA Assessment	CADB Classification
VFN	0.052 / 0.049	0.152 / 0.162	0.139 / 0.142	-
VEN	0.084 / 0.082	0.410 / 0.428	0.232 / 0.241	-
AutoPhoto	0.065 / 0.079	0.407 / 0.427	0.604 / 0.613	-
Q-Align	0.561 / 0.557	0.169 / 0.178	0.809 / 0.804	-
Qwen2.5-VL-32B	0.420 / 0.426	0.195 / 0.205	0.527 / 0.492	0.101
Our Model (7B)	0.763 / 0.777	0.795 / 0.805	0.825 / 0.828	0.583

Citation

If you find our work useful for your research and applications, please cite using the BibTeX:

@inproceedings{photoframer,
    title={PhotoFramer: Multi-modal Image Composition Instruction},
    author={You, Zhiyuan and Wang, Ke and Zhang, He and Cai, Xin and Gu, Jinjin and Xue, Tianfan and Dong, Chao and Zhang, Zhoutong},
    booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition},
    year={2026}
}

Task: Image-Text-to-Text
Type: qwen2_5_vl
License: mit
Languages: en
Created: Apr 17, 2026
Updated: Apr 17, 2026