├── figs
    └── HOG-Diff-framework.png
└── readme.md


/figs/HOG-Diff-framework.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Yiminghh/HOG-Diff/c44dbc9cb079e7d8f140528458971bd5e4db6eb8/figs/HOG-Diff-framework.png


--------------------------------------------------------------------------------
/readme.md:
--------------------------------------------------------------------------------
 1 | <h1 align="center"> HOG-Diff: Higher-Order Guided Diffusion for Graph Generation </h1>
 2 | 
 3 | 
 4 | 
 5 | 
 6 | Official Implementations of HOG-Diff. 
 7 | In this work, we propose a novel Higher-order Guided
 8 | Diffusion (HOG-Diff) model that follows a coarse-to-fine generation curriculum and is guided by higher-order information, enabling the progressive generation of authentic graphs with inherent topological structures.
 9 | 
10 | 🚀 **We are continuously updating this repository and will release the complete code after the paper is accepted.**
11 | 
12 | 
13 | <div align="center">
14 |     <img src="figs/HOG-Diff-framework.png" alt="Framework">
15 | </div>
16 | 
17 | ---
18 | 
19 | 
20 | 
21 | 
22 | ## Environment Setup
23 | This code was tested with PyTorch 2.0.0, cuda 11.8 and torch_geometrics 2.6.1
24 | 
25 | 1️⃣ Download anaconda/miniconda if needed
26 | 
27 | 2️⃣ Create and Activate a Python Virtual Environment
28 | ```
29 | conda create -n HoGD_py39 python=3.9
30 | conda activate HoGD_py39
31 | ```
32 | 3️⃣ Install Dependencies
33 | 
34 | Use the provided requirements.txt file to install all dependencies:
35 | ```
36 | pip install -r requirements.txt
37 | ```
38 | 
39 | 4️⃣ Compile the ORCA Program (for Graph Generation Evaluation)
40 | 
41 | For evaluating generic graph generation tasks, compile the [ORCA](http://www.biolab.si/supp/orca/orca.html) program by running the following command:
42 | 
43 | ```sh
44 | cd evaluation/orca 
45 | g++ -O2 -std=c++11 -o orca orca.cpp
46 | ```
47 | 
48 | 
49 | 
50 | # Running Experiments
51 | 
52 | To train and sample graphs using HOG-Diff, use the following commands:
53 | 
54 | 
55 | ```shell
56 | CUDA_VISIBLE_DEVICES=0 python main.py --config config_name --mode train_ho
57 | CUDA_VISIBLE_DEVICES=0 python main.py --config config_name --mode train_OU
58 | CUDA_VISIBLE_DEVICES=0 python main.py --config config_name --mode sample
59 | ```
60 | 
61 | Replace config_name with the appropriate configuration file.
62 | 
63 | # Citation
64 | Please cite our work if you find our code/paper is useful to your work. :
65 | ```latex
66 | 
67 | ```
68 | 
69 | 
70 |  
71 | 🍀 **Thank you for your interest in our work.** 🍀
72 | 
73 | If you have any questions or encounter any issues while using our code, please don't hesitate to raise an issue or reach out to us directly.
74 | 
75 | 


--------------------------------------------------------------------------------