torch.load fails due to weights_only with BRECDataset

[hugo-hamon]

Hi,

I'm encountering an error when trying to load a dataset using BRECDataset. The following exception is raised:

_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
    (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([torch_geometric.data.data.DataEdgeAttr])` or the `torch.serialization.safe_globals([torch_geometric.data.data.DataEdgeAttr])` context manager to allowlist this global if you trust this class/function.

This is due to the new default behavior of torch.load in PyTorch 2.6+, where weights_only=True by default.

To fix this, I propose explicitly setting weights_only=False in brec/dataset.py (line 50), like so:

self.data, self.slices = torch.load(self.processed_paths[0], weights_only=False)

Let me know if this approach looks good — I can open a PR for it.

Best regards,
Hugo