Simplify model card, add teaser, and update CVPR 2026 citation

#2
by bd469 - opened
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. README.md +11 -34
  3. media/spark-teaser.png +3 -0
.gitattributes CHANGED
@@ -37,3 +37,4 @@ RMBG-1.4/results.png filter=lfs diff=lfs merge=lfs -text
37
  RMBG-1.4/t4.png filter=lfs diff=lfs merge=lfs -text
38
  RMBG-1.4/example_input.jpg filter=lfs diff=lfs merge=lfs -text
39
  media/spark-preview.mp4 filter=lfs diff=lfs merge=lfs -text
 
 
37
  RMBG-1.4/t4.png filter=lfs diff=lfs merge=lfs -text
38
  RMBG-1.4/example_input.jpg filter=lfs diff=lfs merge=lfs -text
39
  media/spark-preview.mp4 filter=lfs diff=lfs merge=lfs -text
40
+ media/spark-teaser.png filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -12,9 +12,7 @@ tags:
12
 
13
  <div align="center">
14
 
15
- <h1>SPARK</h1>
16
-
17
- <p><strong>Sim-ready Part-level Articulated Reconstruction<br>with VLM Knowledge</strong></p>
18
 
19
  <p>CVPR 2026 · <strong>Oral</strong></p>
20
 
@@ -26,36 +24,19 @@ tags:
26
 
27
  </div>
28
 
29
- SPARK is a framework for reconstructing articulated 3D objects from a single RGB image.
30
-
31
- <video controls autoplay loop muted playsinline preload="metadata" width="100%" poster="https://huggingface.co/2bidoubi/SPARK/resolve/d9fadf7111cc965bbcfbab8a9e4ebfab61cf7725/media/spark-preview.jpg" src="https://huggingface.co/2bidoubi/SPARK/resolve/d9fadf7111cc965bbcfbab8a9e4ebfab61cf7725/media/spark-preview.mp4"></video>
32
-
33
- *Input images and articulated reconstructions. An 8-second excerpt from the [full project video](https://heyumeng.com/SPARK-web/static/images/SPARK_video.mp4).*
34
-
35
- ## How SPARK works
36
-
37
- The [paper](https://arxiv.org/abs/2512.01629) describes three main stages:
38
 
39
- 1. **VLM guidance** estimates the part hierarchy and motion, with image generation providing reference views.
40
- 2. **Joint geometry generation** combines part-level and global image guidance to reconstruct the parts together.
41
- 3. **Joint refinement** uses differentiable rendering to match the articulated reconstruction to an open-state reference.
42
-
43
- ## Input and output
44
-
45
- - **Input:** one RGB image of an object.
46
- - **Output:** 3D part meshes and a coarse URDF describing their hierarchy and joints, with textured meshes produced by the texturing stage.
47
 
48
  ## What's included
49
 
50
- This repository hosts the checkpoint bundle used by the [SPARK implementation](https://github.com/YumengHe/SPARK), not a dataset or a standalone text-to-image model.
51
-
52
  | Folder | Component |
53
  | --- | --- |
54
  | `PartCrafter/` | 3D part-generation checkpoint loaded by inference |
55
  | `TripoSG/` | Base-model weights used by the training configuration |
56
  | `RMBG-1.4/` | Optional image background removal |
57
 
58
- The current bundle is approximately **15.6 GB**. Training details and evaluations are described in the [paper](https://arxiv.org/abs/2512.01629).
59
 
60
  ## Get started
61
 
@@ -65,11 +46,7 @@ Follow the [installation instructions](https://github.com/YumengHe/SPARK#-instal
65
  python -m src.utils.download_weights
66
  ```
67
 
68
- Weights are saved under `pretrained_weights/`. The implementation also downloads missing checkpoint folders when they are first needed.
69
-
70
- See the [Quick Start](https://github.com/YumengHe/SPARK#-quick-start) for reconstruction and API configuration. The documented setup uses **Python 3.10, PyTorch 2.5.1, and CUDA 12.4**; the released mesh-inference scripts require an NVIDIA CUDA-capable GPU. External image-guidance and texturing services are separate from this checkpoint bundle.
71
-
72
- In the current public release, joint refinement is provided in separate [optimizer scripts](https://github.com/YumengHe/SPARK/tree/main/URDFoptimizer); it is not run automatically by `run.py`.
73
 
74
  ## License
75
 
@@ -78,11 +55,11 @@ The [SPARK code](https://github.com/YumengHe/SPARK/blob/main/LICENSE) is MIT-lic
78
  ## Citation
79
 
80
  ```bibtex
81
- @misc{he2025spark,
82
- title={SPARK: Sim-ready Part-level Articulated Reconstruction with VLM Knowledge},
83
- author={Yumeng He and Ying Jiang and Jiayin Lu and Yin Yang and Chenfanfu Jiang},
84
- year={2025},
85
- eprint={2512.01629},
86
- url={https://arxiv.org/abs/2512.01629}
87
  }
88
  ```
 
12
 
13
  <div align="center">
14
 
15
+ <h1>SPARK: Sim-ready Part-level Articulated Reconstruction with VLM Knowledge</h1>
 
 
16
 
17
  <p>CVPR 2026 · <strong>Oral</strong></p>
18
 
 
24
 
25
  </div>
26
 
27
+ ![SPARK teaser: articulated objects in a simulated scene](https://huggingface.co/2bidoubi/SPARK/resolve/f56018241fabfa92ea8b4b0b4b28cf01eab50154/media/spark-teaser.png)
 
 
 
 
 
 
 
 
28
 
29
+ Model checkpoints for **SPARK**, a framework for articulated 3D reconstruction from a single image. See the [paper](https://arxiv.org/abs/2512.01629) for the method and results, and the [code](https://github.com/YumengHe/SPARK) for the full pipeline.
 
 
 
 
 
 
 
30
 
31
  ## What's included
32
 
 
 
33
  | Folder | Component |
34
  | --- | --- |
35
  | `PartCrafter/` | 3D part-generation checkpoint loaded by inference |
36
  | `TripoSG/` | Base-model weights used by the training configuration |
37
  | `RMBG-1.4/` | Optional image background removal |
38
 
39
+ The full checkpoint bundle is approximately **15.6 GB**.
40
 
41
  ## Get started
42
 
 
46
  python -m src.utils.download_weights
47
  ```
48
 
49
+ Weights are saved under `pretrained_weights/`. Follow the [Quick Start](https://github.com/YumengHe/SPARK#-quick-start) for reconstruction, requirements, and API configuration.
 
 
 
 
50
 
51
  ## License
52
 
 
55
  ## Citation
56
 
57
  ```bibtex
58
+ @inproceedings{he2026spark,
59
+ title={Spark: Sim-ready part-level articulated reconstruction with vlm knowledge},
60
+ author={He, Yumeng and Jiang, Ying and Lu, Jiayin and Yang, Yin and Jiang, Chenfanfu},
61
+ booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
62
+ pages={7233--7243},
63
+ year={2026}
64
  }
65
  ```
media/spark-teaser.png ADDED

Git LFS Details

  • SHA256: ef9d18b8cabcd51ea8e32526f66ab3af554a122b43c26fdba4ae77fd84062cf3
  • Pointer size: 132 Bytes
  • Size of remote file: 4.25 MB