search

FedML-AI / SpreadGNN

SpreadGNN: Serverless Multi-Task Learning Framework for Graph Neural Networks. Accepted to AAAI22.
44 stars 8 forks source link

consult #2

Open Turningl opened 1 year ago

Turningl commented 1 year ago

Hello professor: when I run your code, I have some problem.

  Traceback (most recent call last):
    File "D:/Federated_learning/SpreadGNN_torch/utils/data_loader.py", line 216, in <module>
      load_data(args)
    File "D:/Federated_learning/SpreadGNN_torch/utils/data_loader.py", line 189, in load_data
      train_data_local_dict[client] = data.DataLoader(train_dataset_client, batch_size=1, shuffle=True,
    File "C:\miniconda\envs\pytorch-gpu\lib\site-packages\torch\utils\data\dataloader.py", line 353, in __init__
      sampler = RandomSampler(dataset, generator=generator)  # type: ignore[arg-type]
    File "C:\miniconda\envs\pytorch-gpu\lib\site-packages\torch\utils\data\sampler.py", line 107, in __init__
      raise ValueError("num_samples should be a positive integer "
  ValueError: num_samples should be a positive integer value, but got num_samples=0

image

It means when I use the function create_non_uniform_split(), client will get no dataset. So if this problem appears, the torch function DataLoader doesn't compile. How do you solve this problem?

GG0614 commented 1 year ago

Hello professor: when I run your code, I have some problem.

  Traceback (most recent call last):
    File "D:/Federated_learning/SpreadGNN_torch/utils/data_loader.py", line 216, in <module>
      load_data(args)
    File "D:/Federated_learning/SpreadGNN_torch/utils/data_loader.py", line 189, in load_data
      train_data_local_dict[client] = data.DataLoader(train_dataset_client, batch_size=1, shuffle=True,
    File "C:\miniconda\envs\pytorch-gpu\lib\site-packages\torch\utils\data\dataloader.py", line 353, in __init__
      sampler = RandomSampler(dataset, generator=generator)  # type: ignore[arg-type]
    File "C:\miniconda\envs\pytorch-gpu\lib\site-packages\torch\utils\data\sampler.py", line 107, in __init__
      raise ValueError("num_samples should be a positive integer "
  ValueError: num_samples should be a positive integer value, but got num_samples=0

image

It means when I use the function create_non_uniform_split(), client will get no dataset. So if this problem appears, the torch function DataLoader doesn't compile. How do you solve this problem?

How about now?