dmlc--dgl
Zihao Ye
9a0511c8e9
* upd * damn it * fuck * fuck pylint * fudge * remove some comments about MXNet * upd * upd * damn it * damn it * fuck * fuck * upd * upd * pylint bastard * upd * upd * upd * upd * upd * upd * upd * upd * upd
32 行
924 B
Markdown
32 行
924 B
Markdown
Predict then Propagate: Graph Neural Networks meet Personalized PageRank (APPNP)
|
|
============
|
|
|
|
- Paper link: [Predict then Propagate: Graph Neural Networks meet Personalized PageRank](https://arxiv.org/abs/1810.05997)
|
|
- Author's code repo: [https://github.com/klicperajo/ppnp](https://github.com/klicperajo/ppnp).
|
|
|
|
Dependencies
|
|
------------
|
|
- MXNET 1.5+
|
|
- requests
|
|
|
|
``bash
|
|
pip install torch requests
|
|
``
|
|
|
|
Code
|
|
-----
|
|
The folder contains an implementation of APPNP (`appnp.py`).
|
|
|
|
Results
|
|
-------
|
|
|
|
Run with following (available dataset: "cora", "citeseer", "pubmed")
|
|
```bash
|
|
DGLBACKEND=mxnet python3 train.py --dataset cora --gpu 0
|
|
```
|
|
|
|
* cora: 0.8370 (paper: 0.850)
|
|
* citeseer: 0.713 (paper: 0.757)
|
|
* pubmed: 0.798 (paper: 0.797)
|
|
|
|
Experiments were done on dgl datasets (GCN settings) which are different from those used in the original implementation. (discrepancies are detailed in experimental section of the original paper) |