dmlc--dgl
Shaked Brody
e2f33fd5cc
* [Model][Core] GATv2 * lint * gatv2conv.py * lint * lint * style and docs * lint * gatv2conv fix Co-authored-by: Shaked Brody shakedbr@campus.technion.ac.il <shakedbr@tangerine.cslcs.technion.ac.il> Co-authored-by: Mufei Li <mufeili1996@gmail.com>
40 行
812 B
Markdown
40 行
812 B
Markdown
Graph Attention Networks v2 (GATv2)
|
|
============
|
|
|
|
- Paper link: [How Attentive are Graph Attention Networks?](https://arxiv.org/pdf/2105.14491.pdf)
|
|
- Author's code repo: [https://github.com/tech-srl/how_attentive_are_gats](https://github.com/tech-srl/how_attentive_are_gats).
|
|
- Annotated implemetnation: [https://nn.labml.ai/graphs/gatv2/index.html]
|
|
|
|
Dependencies
|
|
------------
|
|
- torch
|
|
- requests
|
|
- sklearn
|
|
|
|
How to run
|
|
----------
|
|
|
|
Run with following:
|
|
|
|
```bash
|
|
python3 train.py --dataset=cora
|
|
```
|
|
|
|
```bash
|
|
python3 train.py --dataset=citeseer
|
|
```
|
|
|
|
```bash
|
|
python3 train.py --dataset=pubmed
|
|
```
|
|
|
|
Results
|
|
-------
|
|
|
|
| Dataset | Test Accuracy |
|
|
| -------- | ------------- |
|
|
| Cora | 82.10 |
|
|
| Citeseer | 70.00 |
|
|
| Pubmed | 77.2 |
|
|
|
|
* All the accuracy numbers are obtained after 200 epochs. |