README.md 1.44 KB
Newer Older
1
2
3
4
Graph Attention Networks (GAT)
============

- Paper link: [https://arxiv.org/abs/1710.10903](https://arxiv.org/abs/1710.10903)
5
- Author's code repo (in Tensorflow):
6
  [https://github.com/PetarV-/GAT](https://github.com/PetarV-/GAT).
7
8
- Popular pytorch implementation:
  [https://github.com/Diego999/pyGAT](https://github.com/Diego999/pyGAT).
9

10
Dependencies
11
12
13
------------
- torch v1.0: the autograd support for sparse mm is only available in v1.0.
- requests
14
- sklearn
15

16
17
18
19
20
21
22
23
24
25
```bash
pip install torch==1.0.0 requests
```

How to run
----------

Run with following:

```bash
Chao Ma's avatar
Chao Ma committed
26
python3 train.py --dataset=cora --gpu=0
27
```
28
29

```bash
Chao Ma's avatar
Chao Ma committed
30
python3 train.py --dataset=citeseer --gpu=0
31
```
32
33

```bash
Chao Ma's avatar
Chao Ma committed
34
python3 train.py --dataset=pubmed --gpu=0 --num-out-heads=8 --weight-decay=0.001
35
36
```

37
```bash
Chao Ma's avatar
Chao Ma committed
38
python3 train_ppi.py --gpu=0
39
40
```

41
42
43
44
45
Results
-------

| Dataset | Test Accuracy | Time(s) | Baseline#1 times(s) | Baseline#2 times(s) |
| ------- | ------------- | ------- | ------------------- | ------------------- |
46
47
48
| Cora | 84.0% | 0.0113 | 0.0982 (**8.7x**) | 0.0424 (**3.8x**) |
| Citeseer | 70.7% | 0.0111 | n/a | n/a |
| Pubmed | 78.0% | 0.0115 | n/a | n/a |
49
50
51
52
53
54

* All the accuracy numbers are obtained after 300 epochs.
* The time measures how long it takes to train one epoch.
* All time is measured on EC2 p3.2xlarge instance w/ V100 GPU.
* Baseline#1: [https://github.com/PetarV-/GAT](https://github.com/PetarV-/GAT).
* Baseline#2: [https://github.com/Diego999/pyGAT](https://github.com/Diego999/pyGAT).