dataset load is not successful: shows error

[mmrahman21]

Dear Team,

I was trying to download your HCPGender dataset using neurograph package. I just executed the following code:

from NeuroGraph.datasets import NeuroGraphDataset
dataset = NeuroGraphDataset(root="data/", name= "HCPGender")
print(dataset.num_classes)
print(dataset.num_features)

The dataset was not loaded and shows error. The error is related to PyTorch 2.6 and the weights_only=True change. I am using PyTorch 2.6.

I received this error:

"(1) In PyTorch 2.6, we changed the default value of the weights_only argument in torch.load from False to True. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
(2) Alternatively, to load with weights_only=True please check the recommended steps in the following error message.
WeightsUnpickler error: Unsupported global: GLOBAL torch_geometric.data.data.DataEdgeAttr was not an allowed global by default. Please use torch.serialization.add_safe_globals([DataEdgeAttr]) or the torch.serialization.safe_globals([DataEdgeAttr]) context manager to allowlist this global if you trust this class/function."

If you could take appropriate steps to solve the issue or provide me some guidelines so that I can download the dataset that would be helpful.

Thank you.

Regards,
Mahfuz

---

Md Mahfuzur Rahman
Lecturer, Computer Science
Georgia State University

[Anwar-Said]

Thank you for bringing this to our attention. We will revise the class and submit it to PyG as soon as possible. In the meantime, you may use an older version of PyTorch to download and work with the dataset. I’ll comment here once the class has been revised. Thanks.

[marshuni]

Dear Team,

I was trying to download your HCPGender dataset using neurograph package. I just executed the following code:

from NeuroGraph.datasets import NeuroGraphDataset
dataset = NeuroGraphDataset(root="data/", name= "HCPGender")
print(dataset.num_classes)
print(dataset.num_features)

The dataset was not loaded and shows error. The error is related to PyTorch 2.6 and the weights_only=True change. I am using PyTorch 2.6.

I received this error:

"(1) In PyTorch 2.6, we changed the default value of the weights_only argument in torch.load from False to True. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source. (2) Alternatively, to load with weights_only=True please check the recommended steps in the following error message. WeightsUnpickler error: Unsupported global: GLOBAL torch_geometric.data.data.DataEdgeAttr was not an allowed global by default. Please use torch.serialization.add_safe_globals([DataEdgeAttr]) or the torch.serialization.safe_globals([DataEdgeAttr]) context manager to allowlist this global if you trust this class/function."

If you could take appropriate steps to solve the issue or provide me some guidelines so that I can download the dataset that would be helpful.

Thank you.

Regards, Mahfuz

Md Mahfuzur Rahman Lecturer, Computer Science Georgia State University

I encountered the same issue, and my solution was to add the parameter weights_only=False to all torch.load() methods in the site-packages/NeuroGraph/datasets.py file. Although this approach is not very standard and poses security risks, it does work.