README.md

# Benchmark on Large-Scale Graphs

```
python main.py model=pna dataset=flickr root=/tmp/datasets device=0 log_every=1
```

You can choose between the following models and datasets:

* **Models:** `gcn`, `gcn2`, `pna`, `pna_jk`
* **Datasets:** `reddit`, `ppi`, `flickr`, `yelp`, `arxiv`, `products`