Hugging Face's logo

SynLayers
/
synlayers

Image-to-Image
Diffusers
Safetensors
image-decomposition
layered-image-editing
diffusion
flux
lora
transparent-rgba
Model card Files
xet
 Community
SynLayers commited on
Commit
a5975f0
·
verified ·
1 Parent(s): 3f446ae

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +17 -16
README.md CHANGED
@@ -1,16 +1,18 @@
1
  ---
 
2
  tags:
3
  - diffusion
4
  - image-decomposition
5
  - rgba
6
  - layered-image-editing
 
7
  ---
8
 
9
- # SynLayers Stage 2
10
 
11
- This repository contains the Stage 2 checkpoints and runtime assets for SynLayers, our real-world image layer decomposition system.
12
 
13
- The main assets in this repo are:
14
 
15
  - `SynLayers_checkpoints/FLUX.1-dev`
16
  - `SynLayers_checkpoints/FLUX.1-dev-Controlnet-Inpainting-Alpha`
@@ -19,29 +21,28 @@ The main assets in this repo are:
19
  - `ckpt/pre_trained_LoRA`
20
  - `ckpt/prism_ft_LoRA`
21
 
22
- These assets are used by the public SynLayers demo:
23
  [SynLayers/synlayers](https://huggingface.co/spaces/SynLayers/synlayers)
24
 
25
- The full pipeline uses:
26
 
27
- 1. Stage 1 bbox + whole-caption prediction from [`SynLayers/Bbox-caption-8b`](https://huggingface.co/SynLayers/Bbox-caption-8b)
28
- 2. Stage 2 layer decomposition from this repository
29
 
30
- This repo is intended for the SynLayers decomposition pipeline rather than a single generic `DiffusionPipeline(prompt)` model.
31
 
32
- If you want to see more details of our implementation, please check our paper:
33
  [https://arxiv.org/abs/2605.15167](https://arxiv.org/abs/2605.15167)
34
 
35
  If you find our work useful, please consider citing:
36
 
37
  ```bibtex
38
- @misc{wu2026doessyntheticlayereddesign,
39
  title={Does Synthetic Layered Design Data Benefit Layered Design Decomposition?},
40
- author={Kam Man Wu and Haolin Yang and Qingyu Chen and Yihu Tang and Jingye Chen and Qifeng Chen},
41
- year={2026},
42
- eprint={2605.15167},
43
- archivePrefix={arXiv},
44
- primaryClass={cs.CV},
45
- url={https://arxiv.org/abs/2605.15167},
46
  }
47
  ```
 
 
 
1
  ---
2
+ library_name: diffusers
3
  tags:
4
  - diffusion
5
  - image-decomposition
6
  - rgba
7
  - layered-image-editing
8
+ - synlayers
9
  ---
10
 
11
+ # SynLayers Stage 2 Checkpoints
12
 
13
+ This repository hosts the **Stage 2 checkpoints and runtime assets** for SynLayers, our real-world image layer decomposition system.
14
 
15
+ The main assets in this repo include:
16
 
17
  - `SynLayers_checkpoints/FLUX.1-dev`
18
  - `SynLayers_checkpoints/FLUX.1-dev-Controlnet-Inpainting-Alpha`
 
21
  - `ckpt/pre_trained_LoRA`
22
  - `ckpt/prism_ft_LoRA`
23
 
24
+ These assets are used by our public Space:
25
  [SynLayers/synlayers](https://huggingface.co/spaces/SynLayers/synlayers)
26
 
27
+ The full SynLayers system has two stages:
28
 
29
+ 1. bbox + whole-caption prediction from [`SynLayers/Bbox-caption-8b`](https://huggingface.co/SynLayers/Bbox-caption-8b)
30
+ 2. layer decomposition into transparent RGBA outputs using this repository
31
 
32
+ This repository is intended for the SynLayers decomposition pipeline. It is not meant to be loaded as a single generic `DiffusionPipeline(prompt)` model.
33
 
34
+ For more details, please check our paper:
35
  [https://arxiv.org/abs/2605.15167](https://arxiv.org/abs/2605.15167)
36
 
37
  If you find our work useful, please consider citing:
38
 
39
  ```bibtex
40
+ @article{wu2026does,
41
  title={Does Synthetic Layered Design Data Benefit Layered Design Decomposition?},
42
+ author={Wu, Kam Man and Yang, Haolin and Chen, Qingyu and Tang, Yihu and Chen, Jingye and Chen, Qifeng},
43
+ journal={arXiv preprint arXiv:2605.15167},
44
+ year={2026}
 
 
 
45
  }
46
  ```
47
+
48
+ Thanks for trying SynLayers.