Improve model card for PhysicEdit

#1
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +45 -3
README.md CHANGED
@@ -1,3 +1,45 @@
1
- ---
2
- license: mit
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ pipeline_tag: image-to-image
4
+ tags:
5
+ - physics-aware
6
+ - image-editing
7
+ ---
8
+
9
+ # PhysicEdit: Physics-Aware Image Editing with Latent Transition Priors
10
+
11
+ PhysicEdit is an end-to-end framework for physics-aware image editing that reformulates the editing process as predictive physical state transitions. By combining a frozen Qwen2.5-VL for physically grounded reasoning with learnable transition queries, it provides visual guidance to a diffusion backbone to ensure results are both semantically aligned and physically plausible (e.g., handling refraction or material deformation).
12
+
13
+ - **Paper:** [From Statics to Dynamics: Physics-Aware Image Editing with Latent Transition Priors](https://arxiv.org/abs/2602.21778)
14
+ - **Project Page:** [https://liangbingzhao.github.io/statics2dynamics/](https://liangbingzhao.github.io/statics2dynamics/)
15
+ - **Repository:** [https://github.com/liangbingzhao/PhysicEdit](https://github.com/liangbingzhao/PhysicEdit)
16
+
17
+ ## Usage
18
+
19
+ To use this model, please follow the installation instructions in the [official repository](https://github.com/liangbingzhao/PhysicEdit). You can perform inference on a single image using the provided validation script:
20
+
21
+ ```bash
22
+ python scripts/inference/validate.py \
23
+ --prompt "your_editing_instruction" \
24
+ --image_path /path/to/input.jpg \
25
+ --save_path /path/to/output.jpg \
26
+ --base_model_path /path/to/Qwen-Image-Edit-2509 \
27
+ --dinov2_path /path/to/DINOv2-with-registers-base \
28
+ --lora_path /path/to/physicedit_checkpoint.safetensors
29
+ ```
30
+
31
+ ## Citation
32
+
33
+ If you find this work useful for your research, please cite the following BibTeX:
34
+
35
+ ```bibtex
36
+ @misc{zhao2026staticsdynamicsphysicsawareimage,
37
+ title={From Statics to Dynamics: Physics-Aware Image Editing with Latent Transition Priors},
38
+ author={Liangbing Zhao and Le Zhuo and Sayak Paul and Hongsheng Li and Mohamed Elhoseiny},
39
+ year={2026},
40
+ eprint={2602.21778},
41
+ archivePrefix={arXiv},
42
+ primaryClass={cs.CV},
43
+ url={https://arxiv.org/abs/2602.21778},
44
+ }
45
+ ```