blackhalo-drake / sagnn-substructure-aware-graph-neural-networks Goto Github PK

Traceback (most recent call last):
  File "/home/lugewei/anaconda3/envs/lgw/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/lugewei/anaconda3/envs/lgw/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/home/lugewei/.virtualenvs/untitled2/SAGNN/train/zinc.py", line 100, in <module>
    run(cfg, create_dataset, create_model, train, test)
  File "/home/lugewei/.virtualenvs/untitled2/SAGNN/core/train_helper.py", line 124, in run
    train_dataset, val_dataset, test_dataset = create_dataset(cfg)
  File "/home/lugewei/.virtualenvs/untitled2/SAGNN/train/zinc.py", line 37, in create_dataset
    calculate_stats(train_dataset)
  File "/home/lugewei/.virtualenvs/untitled2/SAGNN/core/utils.py", line 4, in calculate_stats
    ave_num_nodes = np.array([g.num_nodes for g in dataset]).mean()
  File "/home/lugewei/.virtualenvs/untitled2/SAGNN/core/utils.py", line 4, in <listcomp>
    ave_num_nodes = np.array([g.num_nodes for g in dataset]).mean()
AttributeError: 'collections.defaultdict' object has no attribute 'num_nodes'