it cost long time for atom3d/examples/ppi/gnn /data.py

[bozhenhhu]

when run the train.py for ppi/gnn, it spends a lot of time to load data in "for graph1, graph2 in dataloader:", but nothing outputs.

if name=="main":
from tqdm import tqdm

dataset = PPIDataset(os.path.join('/usr/data/data/DIPS-split/data/', 'train'))
dataloader = DataLoader(dataset, batch_size=3, shuffle=False, collate_fn=CollaterPPI(batch_size=3), num_workers=4)
for graph1, graph2 in dataloader:
    print(graph1)
    print(graph2)
    break

is there problems?