search

rusty1s / pyg_autoscale

Implementation of "GNNAutoScale: Scalable and Expressive Graph Neural Networks via Historical Embeddings" in PyTorch
http://arxiv.org/abs/2106.05609
MIT License
158 stars 27 forks source link

synchronize in pool.py #9

Open ghqing0310 opened 2 years ago

ghqing0310 commented 2 years ago

Hello, GNNAutoScale is an interesting and solid work, and thanks for your hard work. However, when running "synchronize = torch.ops.torch_geometric_autoscale.synchronize", it reports an error. How to deal with this problem? image

rusty1s commented 2 years ago

Did you run python setup.py install before executing the scripts?

ghqing0310 commented 2 years ago

Did you run python setup.py install before executing the scripts?

Yes, and when running small_benchmark/main.py, it shows that a circular import happens. image

rusty1s commented 2 years ago

Can you try to run the script not within the root folder, but within small_benchmark?

cd small_benchmark
python main.py
ghqing0310 commented 2 years ago

Can you try to run the script not within the root folder, but within small_benchmark?

cd small_benchmark
python main.py

Yep that's the way I run it. image

rusty1s commented 2 years ago

This is weird. What's the installation log of

python setup.py clean && python setup.py develop